A problem lands on your desk, and before you have thought about it, the AI chat is already open. Many of us do it on impulse, again and again. Over time, that habit determines whether we lead AI, or AI leads us. A 2026 study suggests that opening AI first may shape our thinking in ways we don’t see.

Technology is here for everyone, and staying away from it means falling behind. However, every tool gives us room to grow and takes a few things back. Neil Postman called this a trade-off. I can speak about this from my own experience, which I’m sure many of you have also gone through. In my first job in India in 2003, I had no mobile phone, but most phone numbers which I wanted to remember just stayed with me. Now, with a smartphone, I don’t remember most phone numbers. Other skills took their place, and we grew a lot.

AI is a different kind of trade. My phone holds the numbers. AI can hold our thinking. That is where we need to be careful.

In my last post, You Are the Watcher, I wrote about watching our impulses and the small gap before we act. With AI, that gap is the moment before you type the question. Who should think first?

Mina Lee of the University of Chicago and colleagues asked 393 people to play a city council member. Each read seven documents about a company that had polluted the water, then argued for or against its deal. Some had AI from the start, others later or never. Some got 10 minutes, others 30.

The answer depended on time. When people were rushed, starting with AI helped. With 30 minutes, those who started on their own, with AI later or not at all, tended to write stronger arguments than the AI-first groups. The difference was small.

Few people copied directly from the AI. Still, most of their essays echoed its points. The authors suggest that opening with AI may steer the rest of the essay.

Essay scores rose and fell with time and timing, but people’s ratings of their own critical thinking hardly moved. We may not notice it happening.

In Plato’s Phaedrus, Socrates tells a story set in Egypt. The god Theuth offers King Thamus an invention, writing, to improve people’s memory. People who trust marks on a page, the king replies, will stop using their own memory. They will have “the show of wisdom without the reality.” Postman said he had a point, yet writing gave back far more. AI can too, if we keep thinking for ourselves.

The experiment was one task in one sitting, with small groups. This study does not prove everything, but it is a useful warning.

AI is not bad. It’s very powerful, and sometimes you should absolutely use it early. However, when you have time, and you should make time, your own thinking should come first. Our habits run for years. Day after day, starting from AI’s first answer can slowly wear down our curiosity and judgment.

Before I begin a new project with Claude, I create the whole architecture in my head. I think about where the input comes from and what the output looks like. My first message is usually “Let’s brainstorm first.” I am not technical and don’t understand code, but I understand the business and the logic. So I ask it to walk me through it step by step. Later, I let it run some tasks on its own, but I want every decision to come back to me. It’s a co-worker for me. I’m still the one leading, though.

Think first, then ask. Next time a problem lands, pause before you open the chat and ask, “What do I already think?”

We reach for the phone too, and often no buzz comes first. More on that next time.

Forwarded to you? Then ask our usual question. Was this useful? If it was, subscribe free, and the next one comes straight to you.

P.S. If you like what we do here and find it meaningful, would you invite one person to the discussion? Forward this post, or tap Share below.

Share

Subscribe now

Why am I writing again after a long break, and under a new name? Because AI, which is arguably one of the biggest technological shifts of our time, could become the biggest booster for our growth or could drag us down. It all depends on how we use it. I changed the name from Binary Circuit to The Useful Pause to reflect what I will be writing about going forward.

AI is already reaching people faster than the internet or the personal computer did. By August 2024, nearly 40 percent of Americans aged 18 to 64 had used generative AI.

I believe this is good news. AI can help us learn faster, work better, create more, and solve problems that once took far more time. We should use it to grow.

But we should not hand over our thinking to it.

AI should be a tool we use, not a tool that starts using us. If we stop questioning answers, thinking through problems, or following our own curiosity, we risk losing the judgment that helps us decide what is true, useful, and worth doing.

The same is true beyond AI. Social media, short videos, breaking-news alerts, market apps, and endless notifications all compete for our attention. Technology is not bad. It has made life better in many ways. The real question is whether we are using it with intention.

How do you know when to pause?

For me, it starts with watching myself. In Advaita Vedanta, there is the idea of sakshi, or the witness, the awareness that watches your thoughts and actions without getting carried away by them.

Think about watching a movie. At some point, you can become so absorbed that you feel as though you are living inside the story. Yet you are still sitting in the theater. In the same way, we can get pulled into screens, habits, impulses, and distractions without noticing it.

Being the watcher does not mean withdrawing from life or technology. It means creating a small gap before acting. That gap is where judgment lives.

Before opening an AI tool, checking the news again, or scrolling through another set of videos, I try to ask myself, Is this useful right now? Am I gaining something, or am I just reacting?

I used to check stock prices repeatedly during the day. Now, eight out of ten times, I remind myself, I am not a day trader. Why am I looking at this again? One check in the evening is often enough.

It works on small things too. I like desserts. Every night I would want one. I ask myself, do I really need this? Sometimes I put it aside, or just take one bite. Guess what, most nights I still end up having one. But now I know I am choosing it.

I use AI often. I use it to research, write, learn, and improve my work. But I do not want it to replace the part of me that thinks, questions, and decides.

I am not saying that you stop using it, or that you have to be perfect. Just notice what you are doing and why.

Use AI to grow. Use technology to improve your life. But know when to pause.

You are the watcher.

As a parent of a teen and another child who will be one soon, the subject hits home for me. We worry about the influence of social media on kids.

Some want to ban it for anyone under 16. I’m not against the ban. However, does controlling access solve the problem? When something is off-limits, we often want it more. Prohibition rarely works. Instead, it takes them to something more appealing where they get attention. This slowly becomes a problem, as when they don’t receive attention (likes), the real issue emerges. Phones or social media are not the only problem. Parenting is where it starts.

Follow on LinkedIn

The conversation often turns to restricting access. But do you think control is really the solution, or the first line of defense? The truth is, prohibition can often make something more appealing. I believe the real challenge isn't with the technology itself, but in how we parent.

We often give kids devices to occupy them during our own busy moments—at the gym, during dinner, or with friends. Over time, that device becomes their sanctuary.

We, as parents, will have to take the first responsibility and accountability. Instead of building walls, we need to create awareness. We need to be more involved in their lives. Listening to our children as much as possible, rather than telling them what to do, will go a long way.

This is a work in progress for my family, but I trust we're moving in the right direction by focusing on conversation and understanding rather than just rules.

What do you think? Are we ready to parent intentionally in the digital age?

Leave a comment

Schedule a discovery call

Welcome to Binary Circuits’ 9^th weekly edition

This week's must-know developments:

• Great AI debate: Raw Computational Power vs. Efficiency. Who wins?

• Virtualization is imperative: AR/VR is taking a leap forward!

• Overwhelmed with Innovation? Steps to navigating the AI tsunami.

Let’s dive in!

Binary Circuit is a reader-supported publication.
To receive new posts and support our work,
consider becoming a subscriber.

Subscribe now

More computational power vs. engineering and algorithmic efficiency. The race to build superior AI models. What’s more important?

A simple equation has long driven the AI race: more computing = better models. DeepSeek-V3 questions this hypothesis. The DeepSeek-V3 model proved that improved algorithms and efficiency could be just as crucial as actual computer capacity.

DeepSeek-V3 attained competitive performance without the same degree of computational capability as Western leaders, including OpenAI, Anthropic, and Google DeepMind. DeepSeek emphasizes more efficient architectures that maximize how models process data, better training strategies to squeeze more performance from fewer resources, and intelligent use to boost learning efficiency instead of depending on brute-force scaling.

This begs a troubling issue for Western AI companies: what if more computation isn't necessarily the solution? Is scaling computing still the dominant strategy, or are we reaching a turning point where algorithmic improvements matter more?

We believe that while efficiency gains are making strides and will play a significant role, the reality is that no one has enough computing yet to realize AI's potential fully.

Case in point: Elon Musk, a genius of our times in optimization and breakthrough technologies, is betting on more computational power.

Elon Musk’s xAI is building one of the most powerful AI supercomputers in the world—the Colossus Computer. Musk believes that AI models are fundamentally constrained by computing power, and to stay competitive with OpenAI, Google, and DeepMind, xAI needs massive GPU resources to train Grok-3 and beyond.

His argument? Even if efficiency gains matter, the most significant breakthroughs will still require unprecedented computing power. We tend to agree based on the evidence.

The Colossus Computer is an engineering marvel. Musk and his team took an unprecedented short period of 122 days to install 100,000 H100s and then added the same number in just 92 days. The target is to install one million GPUs. Musk started training Grok-3 as soon as they installed the first GPU rack on the 19^th day, a feat that only he could have pulled off.

Colossus utilizes advanced liquid cooling solutions, including cold plates and coolant distribution units, to manage the immense heat generated by the GPUs. It's the largest liquid-cooled GPU cluster deployment in the world.

The result: xAI launched Grok-3 this week, which used 200 million GPU hours for training—10x-15x more than its predecessor, Grok-2. Elon Musk referred to Grok-3 as “scary smart.”

Key data points about Grok-3:

• Benchmark Performance: Early reports suggest that Grok-3 outperforms leading models like OpenAI's GPT-4o and DeepSeek's V3 in areas such as mathematics, science, and coding, though independent verification is pending.

• Enhanced Reasoning Capabilities: Incorporates advanced reasoning modes, such as "Think" and "Big Brain," enabling the model to tackle complex, multi-step problems by allocating additional computational resources.

• Deep Search Functionality: Introduces a next-generation search engine that retrieves information and provides detailed explanations of its reasoning process, enhancing transparency and user trust.

However, it’s not just the brute force. xAI has improved Grok 3’s capabilities beyond increased computing power. xAI incorporated:

• Synthetic datasets, artificially generated, simulate various scenarios and improve learning efficiency while addressing data privacy concerns.

• Self-correction mechanisms allow models to identify and correct mistakes by comparing outputs with known correct responses, reduce errors, and improve accuracy.

• Reinforcement learning, a type of machine learning, trains models to maximize positive outcomes through trial and error, enhancing decision-making.

xAI also integrated human feedback loops and contextual training to reduce hallucinations and produce natural human-like responses. Contextual training teaches AI to understand and adapt responses based on context, considering previous interactions, user intent, and surrounding information.

So, what’s more important: computing power or efficiency?

Our take: Compute is king, but efficiency will gain importance as scaling laws slow down the progress through brute force. While companies like DeepSeek demonstrate that efficiency gains can bridge the gap, no one in the AI sector believes they have “enough” computing yet.

The contest between compute-intensive and efficiency-first AI strategies is just beginning.

Why should you keep an eye on the rapid progress of AR/VR technology?

AR and VR technologies are undergoing a massive development cycle. Businesses like Meta, Google, and Samsung are heavily funding the accessibility and practicality of these technologies.

Meta is dominating the AR glasses space with its prototype Orion eyewear. Orion is lightweight and less obtrusive than typical VR headsets, making it suited for daily use. Experts expect that AR devices will become more glasses-like. This transition is vital for broader acceptance as devices become less bulky and maybe fashionable.

Ray-Ban Meta Smart Glasses sold one million units in 2024, demonstrating AI-driven wearable demand. Meta keeps adding AI to Ray-Ban smart eyewear. These upgrades include object identification, live language translation, and music streaming service integration.

By 2025, Meta intends to have cumulatively spent more than $100 billion on VR and AR. Its AR/VR unit, Reality Labs, invested $19.9 billion in 2024 and another $20 billion in 2025. Meta is dedicated to making AR and VR a widely usable technology even though late 2024 results show $5 billion lost every quarter.

Project Moohan from Google and Samsung is notable.

To challenge Meta, Google and Samsung are unveiling Project Moohan, a mixed-reality helmet scheduled for 2025. Operating on Android XR, this helmet will broaden the AR/VR industry and include AI-powered interactions.

Other smaller firms are also striving to make AR/VR more accessible and adaptable. Beyond the big players, several businesses are improving AR/VR glasses and devices. The lightweight and cheap Xreal Air 2 Ultra focuses on screen mirroring. TCL NXTWEAR S offers a pleasant AR experience with a substantial virtual display for entertainment and casual use.

Why should you care as a business owner?
AR/VR applications will significantly enhance operations in various industries and open avenues for new strategic business models. Examples include:

• Improved Customer Engagement: Companies can design interactive showrooms and online buying environments.

• Remote Work & Training: VR-powered conferences help to enhance learning and communication.

• Healthcare: AR and VR have applications in pain management, surgical procedures, therapy, and healthcare professional training.

• Education: VR and AR enhance learning through immersive environments and interactive elements.

• Marketing and advertising: VR lets organizations create captivating ads and product demos.

AI-powered smart wearables and tools are expected to impact customer and company behavior despite the current challenges for companies like Meta to make a profit. To stay ahead, businesses should consider how AR/VR may improve operations and engagement.

Share

The conundrum facing businesses is negotiating the flood of innovation, especially in AI. What can you do?

The AI development is moving at a breakneck speed. Almost every day, new models, tools, and platforms surface, each seeming to be the next great thing. For business users investigating AI for their growth, it is a cause of mounting concern; it is exciting and overwhelming. Every business right now is grappling with questions like, how can we stay current with this unrelenting rate of change? Where should we spend our time and money? What should we disregard, and what should we adopt?

The challenge is that businesses can’t integrate and validate the applicability of particular AI models as fast as the technology keeps changing. It is best explained through Martec’s law, which states that organizations change logarithmically while technology develops exponentially. This leaves a gap between our realistically manageable capacity and what is possible.

For consumers of artificial intelligence, this discrepancy shows up in numerous forms.

The sheer number of AI tools may paralyze one. Though we research and compare for hours, we often find it challenging to decide. Even if the newest trends don't fit our requirements or priorities, the FOMO Trap—the fear of missing out—drives us to pursue them. We hop from one tool to the next, never really learning any of them. Frustration results from this, and one feels always behind the curve.

How, thus, should we negotiate this AI-powered maze? These tactics are some things to give thought to:

• Consume with awareness—ask yourself, I'm trying to fix what problem? Is this product really addressing that issue, or am I being drawn in by the buzz?
  Can I feasibly include this tool in my current process?

• Emphasize depth above breadth—instead of trying to master multiple tools, focus on a handful that meet your needs and practice them. Much more profit will result from this than surface tinkering.

• Learn to say “no”—though intriguing, declining new technology is acceptable. Time and attention are valuable. Avoid FOMO-driven decisions.

• The Human-AI Cooperation— Remember that AI is a tool, not a substitute for human acumen. The ideal approach is to use AI with your own intelligence.

The AI revolution is altering the world, and it’s natural to get caught up in the excitement. At Greenlight, we empower you to exploit AI’s promise without being overwhelmed by its rapid progress by being careful, strategic, and discriminating.

Chart of the Week:

AI’s scaling has been dependent on how much data, parameters, and computational power can be fed to the models. The charts below from “Our World in Data” demonstrate how stupendous the rise has been.

Since 2010, the training data has roughly doubled every nine to ten months. Datasets for training large language models have grown faster, tripling in size annually since 2010. GPT-2 was released in 2019 after training on 4 billion tokens (or 3 billion words). GPT-4 was trained using approximately 13 trillion tokens, or 9.75 trillion words.

Since 2010, the number of AI model parameters has about doubled annually. Large models like GPT-3 have around 175 billion parameters. Epoch's QMoE model has the highest estimated parameters at 1.6 trillion.

Since 2010, compute growth accelerated, doubling every six months. On average, computing doubled every two years from 1950 to 2010. The most compute-intensive model reaches 50 billion petaFLOP.

GPUs, crucial for training, are becoming increasingly powerful, doubling computer speed every 2.5 years for every dollar spent.

Sound bites you should know:

• BCI becomes two-way. Chinese researchers created the world’s first two-way brain-computer interface (BCI), which is 100x more efficient than existing systems. This breakthrough allows the brain and computer to learn from each other, improving control accuracy and allowing medical and consumer technology.

• AI learns like humans. Researchers have developed a new AI algorithm called Torque Clustering that can learn without human supervision. The algorithm has shown 97.7% accuracy across 1,000 datasets.

• Coding for everyone. Replit, Anthropic, and Google Cloud are using AI-powered coding tools to let non-technical individuals develop software. The system has already built over 100,000 applications.

• Quantum computing is in the spotlight. Microsoft unveils Majorana 1, the world’s first quantum chip, using a “topological core” architecture. Powered by topoconductors, this chip lays a foundation for one million qubits and scalable quantum computing.

• AI buzzwords. What do AI jargons, like reasoning, evals, and synthetic data, mean, and what are their significance? Read more here.

Thank you, for reading! Let us know if you have any feedback. Really appreciate if you could share Binary Circuit and help us grow.

Cheers,

Team, Binary Circuit/Greenlight

Schedule a discovery call

Welcome to Binary Circuits’ 8^th weekly edition

This week's must-know developments:

• AI’s rapid improvement and wider penetration is rewriting the job profiles and business models. Is your business taking note?

• AI is rapidly changing software development, with predictions that it will soon automate most coding tasks. How do you adapt or leverage the change?

• AI's ability to handle complex thinking reflected in improving score on “Humanity’s Last Exam” suggests the technology will be able to perform complex roles. Business agility is important.

• Workplace burnout is becoming a significant concern. What can you do?

Let’s dive in!

AI Reinvents the Very Skill That Built the Digital Age—What’s Next?

Software development has been the pillar of the digital revolution for decades now. It changed whole sectors, started billion-dollar businesses, and was considered one of the safest professions. Parents in my generation in India, where I grew up, had one apparent wish for their kids: learn to code. It was the path to a reliable and sustainably successful career.

Now, generative AI (GAI) is challenging that assurance. AI is accelerating the rate of technical advancement itself. It is like a self-improving machine, constantly developing the instruments and enhancing its power. Think about the ramifications: software, which fueled the digital revolution, is now being rewritten by the revolution’s offspring. 

AI replaces rather than just helps programmers (see section below). Early this year, Salesforce ceased recruiting software engineers. By the end of this year, Mark Zuckerberg predicts AI will be automating mid-level developers. Sam Altman pointed out that by the end of 2025, “software engineering is going to look very different.”

The rate of change is accelerating!

While technology has continually transformed sectors, GAI is increasing the rate of change itself rather than only providing another wave of invention. Once taking decades to develop, these events occur in months.

AI will thus think, decide, and strategize better than most experts, not only assisting professionals.

GAI is non-discriminatory! What sectors, occupations, and skills follow? The trend suggests that information processing, analysis, and creative problem-solving fields may be vulnerable.

Consider law, business, and research. DeepResearch, released by OpenAI and Google (yep, both firms named it the same), shows how research and analytics can be enhanced and reinvented. DeepResearch delivers market insights in minutes that would have cost thousands in consulting fees, beating financial analysts.

But it won’t stop there.

Reinvention is possible in science, engineering, architecture, and product development. AI is no longer only just executing instructions. It is developing solutions that humans would not have considered.

Agentic AI is a game changer. When aggregated in great numbers, AI agents—specialized in particular tasks—form a workflow that can challenge once-considered safe havens, impervious to the kind of automation we linked with factories and assembly lines. The agentic AI era is unfolding in front of us! (more on this next week).

What to make of it? The lesson is obvious for business owners: complacency is an existential risk. The new survival skills include adaptability, acceptance of change, and seeing the changes in technological turmoil. It’s about leveraging GAI, not denying it.

How can you incorporate these fast-developing innovations into your business activities? Think, how many roles and duties will be redefined or augmented by GAI?

The challenge is much higher for career aspirants. The skills you think are essential today could be outdated tomorrow. The secret is not just to study but also to develop adaptability, agility, critical thinking, and a lifetime passion for learning. Emphasize the incredible human abilities of imagination, empathy, and sophisticated problem-solving. These are the domains where, for now, at least, GAI finds it difficult to compete.The world is changing faster than ever before. Foundations are being shaken.

The question isn’t whether disruption is coming; it’s how we will respond. Will the tide carry us away, or will we be the architects of this new planet?

Generative AI is eating software development

Jensen Huang, CEO of NVIDIA, has argued since 2022 that AI will significantly reduce the need for coders. At the annual Abundance 360 Conference 2023, Emad Mostaque, then CEO of Stability AI, predicted that the world would not need software programmers within five years.

Late last year, a panel led by Dr. Peter Diamandis at the FII Summit concluded that software development going forward would be for people passionate about coding. The euphoria of software development being a highly paid skill set is no longer valid with the proliferation of capable generative AI (GAI).

More and more companies are proving this to be true. Google explicitly said that a quarter of all code at Google is AI-generated. GitHub’s Copilot, built on OpenAI’s models, now assists millions of developers, generating 30%–50% of new code in some workflows. Jensen Huang has repeatedly mentioned that NVIDIA’s software programmers and engineers now use AI assistants to help with programming and debugging software.

The Anthropic Economic Index report analyzed 4 million conversations with GAI assistant Claude to understand how the technology is used in the workforce. The study reveals the highest use for tasks in software engineering roles.

Among the GAI-first coding firms that have surfaced, Zencoder, Merly, Cosine, Tessl, and Poolside are all attempting to advance the field of generative coding. In fact, Cursor AI, a GAI-powered code editor, rose the fastest (roughly a year) to $100 million in annualized recurring revenue from $10 million in 2023.

Classic AI coding tools were like autocomplete on steroids, helping developers create cleaner code. The next generation, led by Cosine and Poolside, builds models that think like engineers with logic, reasoning, and creativity.

GAI tools create more than functional code. They analyze development workflows, fix errors, and reason through complex challenges to learn how software is developed.

The result? GAI is designing entire applications. Developers are becoming GAI supervisors.

At Cosine, engineers have been banned from using coding assistants except for their AI tool, Genie. They now spend less time writing code and more time reviewing AI-generated solutions.

This shift is already visible across the industry. Instead of manually coding, developers increasingly orchestrate AI models, feeding them high-level prompts and correcting outputs.

Analogy in perspective: This new workflow resembles ATM banking. Once, thousands of tellers manually processed transactions. Now, software automates the process, and humans only intervene when needed. The same pattern is now playing out in software engineering.

Companies will require fewer programmers and more AI-driven development specialists. Understanding how AI produces software will determine success, not coding.

What happens when GAI can code better than humans?

Please note that AI coding tools still need human monitoring. Though they can create accurate syntax, they sometimes fail at functional correctness—writing programs that perform what you intend.

However, companies like Zencoder and Poolside are making rapid progress to close this gap. GAI is being trained on software development thought processes and finished code.

Their approach? Give AI a breadcrumb trail—capturing human developers’ exact steps when solving problems.

Reinforcement learning is valuable in this context. AI coding models are developing and testing their code thousands of times per second, just like DeepMind’s AlphaZero learned chess and the game of Go by playing against itself.

Why does it matter?

An AI system could soon design a whole application from scratch, determining the architecture, coding the features, troubleshooting faults, and optimizing performance without human involvement.

What is the burnout epidemic, and what should you do about it as a business owner?

Entrepreneurship is credited with sparking innovation, creativity, and economic prosperity. Many people dream of starting from scratch and being their own boss, and governments encourage it. However, like everything, it also comes with a hidden cost—burnout.

Burnout is now a common problem impacting employees in modern society. Jennifer Moss explores the underlying reasons for this epidemic in her book, The Burnout Epidemic: The Rise of Chronic Stress and How We Can Fix It, stressing that conventional self-care benefits are inadequate remedies. Knowing about burnout and putting good plans in place to fight it inside your company is absolutely vital for business owners.

But what is burnout? How would you characterize it, and what are the symptoms?

Burnout is a psychological disorder distinguished by emotional tiredness, depersonalization, and a reduced sense of personal accomplishment. It’s a persistent condition of physical and emotional depletion, not only sporadic stress.

The symptoms are constant fatigue, negative attitude toward employment, poor professional performance, and even health conditions, including gastrointestinal problems or headaches. The timeliness of intervention depends on early recognition of these indicators.

Together with UBS Global Entrepreneur Report 2025 research, Moss’s work emphasizes the actual scope of the issue. According to a recent study, 34% of managers at small and medium enterprises reported burnout, and 87% of them have suffered from poor mental health.

Why is this thus occurring? Burnout transcends the person, as Moss underlines. Often, it results from a toxic workplace, high expectations, lack of control, and a disconnection from the purpose of the work. Constant pressure and an always-on culture help to explain why our job life seems meaningless.

The UBS Global Entrepreneur Report 2025 says entrepreneurs want to blend work and investing with personal interests. Moreover, half (52%) want to prioritize travel, hobbies, and family.

How can a business owner help prevent burnout among their staff?

The burnout epidemic is beyond the standard “self-care benefits.” Those are good but are only band-aid solutions on a far more severe cut.

These tactics are some things to give thought to:

• Culture of psychological safety—establish an environment where staff members may express worries, stress, or workload without fear of consequences. Encourage honest communication and proactive feedback seeking.

• Redefine success—question the always-on attitude. Does everyone really have to be available around the clock? Review success measures. Are you just present or producing satisfying results? According to UBS research, many business owners yearn for a better work-life balance; your staff certainly does.

• Empowerment and autonomy—give staff members more influence over their work. People who feel they have agency are less prone to feeling overwhelmed and powerless.

• Meaningful work—help employees relate to the overall cause they are working toward. Let them become part of the bigger picture. Guide them in seeing how their efforts support the objectives of the business. Work that feels significant is more enjoyable and less taxing.

• Appropriate workloads and expectations—review your task often to ensure it is doable. As necessary, be ready to change goals and timeframes. Don’t exalt diligence.

• Lead by example—you create the tone as a business owner. Give your health a top priority and show a good work ethic. Take pauses, disengage after work, and prioritize family and personal life.

Though it’s a significant problem, burnout is not insurmountable. Changing the emphasis from personal self-care to institutional reforms inside the company will help to produce a better, more engaged workforce. Not only is it the moral thing to do, but it also allows business to grow.

Chart of the Week:

Consider it—just last year, GAI grappled with sophisticated reasoning. Designed to test AI with the most challenging issues from law, philosophy, mathematics, and physics, a group of researchers developed Humanity’s Last Exam (HLE). The HLE consists of 3000 questions, spans various fields requiring diverse cognitive strengths, and challenges Nobel Prize-winning minds and those with Einstein-level IQs.

AI models achieved hardly more than 10% accuracy last year.

Fast forward to now: AI is currently scoring over 26%. If this trend continues, it may surpass human intelligence before the year ends.

Sound bites you should know:

• Tesla’s self-driving fleet is now logging over a million miles per year—navigating complex private zones without human intervention. Is full FSD around the corner? 

• Amazon announced a capex of over $100 billion on AI in 2025 (up from $78 billion in 2024), primarily for AWS. One more example is that bets on increased AI demand will offset lower prices.

• Elon Musk places a $97.4 billion bid for OpenAI’s assets. Sam Altman rejects. However, this could change its nonprofit valuation, pushing Sam Altman and investors to reconsider their AI for-profit investments.

• Anthropic CEO Dario Amodei predicts AI will reach the intelligence of an entire nation within two years. Global governance is moving too slowly to keep up.

• Perplexity Sonar, powered by chipmaker Cerebras, is challenging conventional web search with lightning-fast AI inference and superior answer quality. Google should pay attention!

Schedule a discovery call

Welcome to Binary Circuits’ 7^th weekly edition

This week's must-know developments:

• The fast expansion of AI confronts a question: is the hype exceeding practical uses?

• Race to lead AI development is in hyperdrive. OpenAI, Google, and China's AI behemoths drive invention at an explosive pace.

• Can AI-generated content be copyrighted? A landmark ruling could change everything for creators.

• Leadership traits for the next generation: Companies need leaders with key traits like resilience and humility. Developing future leaders is crucial.

Let’s dive in!

Share

AI landscape is entering a hyper-competitive environment. Is too much money chasing AI with no “killer applications” (yet). How much is too much? Where is this heading? What to watch out this year?

We believe that generative AI is a step change in technological innovation. The technology has the potential to hypercharge several other exponential technologies’ development. In 2024, frontier companies launched several new models with increased capabilities. We saw a glimpse of agentic AI, open-source gained momentum, and reasoning models started to take center stage. The launch of DeepSeek-V3 in December and its reasoning model R1 seem to have put oil on fire.

What’s driving the trend?

Computational performance is multiplying quickly. ARK Invest Big Ideas 2025 put this in the context of a chess board. The idea is that if you start from the first square and double on every square, you end up doing 64 doublings (power of compounding). This means performance per dollar of AI computing power could improve >1000x by 2030.

Secondly, AI training costs are declining at roughly 75% per year, and inference costs at nearly 90%. According to Andreessen Horowitz, LLM inference costs have dropped 1,000x in three years. In November 2021, GPT-3 (MMLU 42) cost $60 per million tokens. Llama 3.2 3B from Together.ai now achieves the same score for just $0.06 per million tokens.

Two main trends are evolving in AI development.

First, the frontier model developers, like OpenAI, are still in the camp to put in as much money as possible and as fast as possible to scale the models and be the first to develop artificial general intelligence (AGI). Last week, Sam Altman said at a conference with Masayoshi Son that he will announce AGI within two years. Depending on how one defines AGI, there are several forecasts.

The second camp is the companies that use the frontier models and apply techniques like distillation and quantization, among others, to develop small language models with niche models capable of doing specific tasks better than others. This is also leading to more use of reasoning models.

The models coming out of China are competitive with the Western companies at lower costs (debatable). The development has created a panic environment and pressure to reduce costs and increase performance. A flurry of models has been announced since the DeepSeek-R1 release on January 20 (see notes below).

The key question is how far the frontier companies will go in terms of investments to remain competitive.

Will the focus turn to developing actually usable apps that generate revenues? Is this creating an environment where too many companies spend a lot of money doing almost the same thing? This might lead to many losers.

We believe that 2025 will be a defining year in the AI landscape. Trump administration in the U.S., with more technological representation in his team than ever, could open more ways for investments and developments while reducing regulations.

China will continue to catch up and maybe supersede the U.S. in some areas of AI development. Our conviction is that AI deployment to generate revenues will gain traction in China much faster than in the U.S.

Agentic AI applications where the content generation will scale up to act will start taking shape, with hopefully some commercial applications.

Model compression to run locally on devices will become the new focus area of R&D, which promises to increase privacy and lower energy footprint (could derail NVIDIA’s train?).

The regulatory debate on whether the energy used by the models should be restricted could become much more visible. Reasoning models use more energy during inference as they think more.

We believe the winners will be the companies in the application layer and, to some extent, the platform layer.

Where do you see AI heading?

The global showdown of the AI models—a race to innovate faster

Since DeepSeek-R1 arrived on January 20, 2025, the AI space has become a high-stakes conflict. DeepSeek V3 really stood out because of its engineering solutions driving efficiency. With the help of reinforcement learning, R1-Zero raised the stakes.

A constant stream of new models boasting superior performance has increased the competition. While OpenAI is under pressure to innovate to maintain its top position, Chinese AI companies fiercely compete among themselves and against global giants.

• OpenAI’s o3-mini, directly responding to DeepSeek’s R1, gives high performance at a competitive price.

• Google joins with the launch of the 2.0 line of models, its most efficient version, and in line with its strategy to lead in AI agents.

• Alibaba’s Qwen 2.5-Max now claims to work better than GPT-4o, DeepSeek-V3, and Llama-3.1-405B.

• Tülu 3 (AI2) is shaking up the open-source world and doing better than DeepSeek-V3 and GPT-4o in many ways.

The number of new models is staggering, but one thing is certain: artificial intelligence is evolving at an unprecedented rate.

Three key trends to note:

Open-Source against Proprietary AI: Who wins? History favors Open Source.

A big battle is going on between fully open and partially transparent models. The open-source movement is supported by Tülu 3 from Ai2 because it makes all of its training data, code, and weights available to everyone. Other top models, on the other hand, are more or closed-ended.

Reinforcement Learning from Verifiable Reward (Tülu 3) is a novel approach encompassing:

• Effective parallel computation among 256 GPUs

• Synchronized weight optimization

• Balanced computational distribution across 32 nodes

• Integrated vLLM using tensor parallelism in sixteen directions

The trifecta of speed, accuracy, and cost in AI performance.

Now included into ChatGPT, OpenAI’s o3-mini shows:

• 24% faster reaction times than o1

• improved scientific and mathematical accuracy

• Pricing $0.55 per million input tokens and $4.40 per million output tokens

Although OpenAI is quickly closing the pricing difference, DeepSeek R1 is still more reasonably priced.

Google’s 2.0 model line-up could be a harbinger of the proliferation of AI agents.

Google announced new models in the 2.0 series this week:

• Gemini 2.0 Flash: The fast, reliable workhorse is now generally available.

• Gemini 2.0 Flash-Lite: A budget-friendly option for broader access.

• Gemini 2.0 Pro: Complex task and coding juggernaut.

Key enhancements and new features include:

• Massive data processing: Gemini 2.0 Pro can process vast volumes of data with a 2-million token context window.

• Integration: Added Google Search and code execution for dynamic apps.

• Multimodality: The model can understand and generate text, graphics, video, and audio.

Google’s enhanced reasoning capabilities improved its benchmark testing scores. It got 73.3% on AIME (math) and 74.2% on GPQA Diamond (science).

Price and accessibility:

• Gemini 2.0 Flash costs 10 cents for every million tokens to enter text, pictures, or videos.

• Flash-Lite costs only $0.0075 per million tokens, making it an even cheaper choice.

Google is using AI tools to make web searches more engaging and rewarding.

CEO Sundar Pichai listed three critical projects:

• Deep Research: AI agents that write detailed search reports.

• Project Mariner: Like OpenAI’s Operator, it browses websites for consumers.

• Project Astra: A real-time AI assistant that can instantaneously process live footage and answer inquiries, possibly enabling AR eyewear.

China’s AI rise – watch out for cost and real-life killer apps.

In 2024, Chinese AI companies only raised $5.2 billion. In contrast, U.S. AI companies raised $76.3 billion. However, technology leverage is helping China improve and fight on the global stage.

China’s AI strategy optimizes strong AI for low-resource hardware, which could shift the game.

On many tests, Alibaba’s Qwen 2.5 beats DeepSeek-V3.

ByteDance, Moonshot AI, Zhipu AI, Baichuan AI, MiniMax, and DeepSeek are aggressively launching models and penetrating global markets.

Many of these “AI tigers” have reached $1B+ valuations, supported by China’s IT giants and government.

AI Agents are beginning to shake the productivity matrix.

From a research tool, artificial intelligence is becoming a regular everyday productivity booster.

• Google’s Gemini Live – Pixel 9 users now have instant access with context-aware AI assistance. It allows users to access information instantly, ask questions, and receive context-aware responses, thereby boosting productivity

• OpenAI’s Operator – Automates browsing, bookings, and online tasks.

• Deep Research (OpenAI) – A novel feature outperforms GPT-4o and o3-mini in expert study. This feature enhances ChatGPT’s research capabilities by allowing users to explore complex topics, gain specialized information, and get detailed answers to complex questions.

Challenging times for Businesses looking to leverage AI – what can they do?

The rapid AI innovation benefits businesses with better pricing and more choices, but it also brings challenges:

• Decision paralysis – Too many options make it hard to choose the right model.

• Tech lock-in risks – Companies struggle to future-proof their AI investments because of the risks of lack of interoperability.

Businesses must evaluate quality, speed, latency, price, and context window size. Platforms like Artificial Analysis, ChatLabs, and ChatBot Arena help compare models and find the best fit.

At Greenlight, we specialize in cutting the noise and tailoring the IT solutions for your specific needs. Book a discovery call to discuss with our network of experts.

Subscribe now

AI content with a human touch can be copyrighted now. What does it mean, and why is it important to know?

A landmark U.S. Copyright Office decision states that AI-created works can be protected by copyright if there is clear evidence of human creativity. This verdict is crucial for content creators, musicians, and artists using AI tools because it establishes ownership in the era of generative AI.

What does the ruling state?

Sole AI creations are not copyrighted. The Copyright Office has underlined that the preservation of copyright depends mostly on human innovation. Copyright may apply if someone alters, organizes, or adapts AI-generated content to reflect their creative effort.

Why does it matter?

This decision dramatically affects businesses that depend on AI for creative work, such as:

• Entertainment & music — AI-assisted compositions and scripts can now be copyrighted, encouraging studios and artists to use AI.

• Digital art & design: Artists using AI technologies like MidJourney or Stable Diffusion may protect their work with enough human originality.

• Marketing and material creation - AI can safeguard branded material against copyright infringement.

However, the ruling does not address the contentious issue of AI training on copyrighted material.

Impacts on businesses and AI development.

It will change how businesses and creators use AI and create new standards for human-AI collaboration:

• AI adoption accelerates - Clearer copyright laws will encourage more enterprises to use AI in creative processes.

• New revenue streams: AI-assisted creations allow artists and creators to create more material and develop their businesses.

• Human-AI cooperation becomes the standard: Focusing on human input will drive organizations to combine AI and creativity for better results.

• Licensing & liability risks: Companies must disclose and use AI-generated material reasonably to avoid legal issues.

The image below from Rose Enigma (more about the artist here) shows the human-AI collaboration.

Even though AI is here to stay, human ingenuity is paramount. As corporations and creators navigate this new landscape, those who embrace human-AI synergy will lead the next wave of media, art, and entertainment development.

Why leaders should focus on developing the next generation of leaders, and how.

Leadership is more complicated than ever. Global risks, rapid technical breakthroughs, and changing labor demands require a fresh approach. Leadership is about developing future leaders, not just navigating complexity.

Many believe leadership implies making decisions, giving commands, and enforcing them. The best leaders inspire rather than command. They create an environment where people want to follow, not have to. In a world where workers wish for autonomy and significance, leadership must shift from control to cooperation, trust, and mentorship.

CEOs must manage talent. They will find and develop leaders. They must promote trust, transparency, and learning. CEOs can ensure their organizations thrive in constant change by exhibiting the desired leadership attributes and building a strong leadership factory.

The McKinsey article noted that this perspective shift required a fundamental leadership rethinking. Leaders are no longer lone heroes making all the decisions. These leaders must be servant leaders who prioritize team success and the organization’s mission. They must be humble, open to learning, and vulnerable. This builds trust and teamwork, which is necessary in an agile and adaptable society.

Cultivating Essential Traits:

A McKinsey article identifies six key traits for 21st-century leaders:

• Positive Energy and Inspiration: Leaders need to radiate optimism and inspire their teams, even in periods of uncertainty.

• Servant Leadership: Focus on empowering others and enabling their success.

• Continuous Learning: Embrace curiosity, take risks, and learn from mistakes.

• Grit and Resilience: Remain steadfast in facing challenges and bounce back from setbacks.

• Cheerfulness: Use humor to build connections and defuse tension.

• Stewardship: View themselves as caretakers of the organization, focused on long-term sustainability and leaving a heritage.

Creating Leadership Factory:

Great leaders produce more leaders. The “leadership factory” notion applies here. It’s about creating an organization-wide leadership culture.

Key Elements of a Leadership Factory:

• Define Leadership Attributes: Clearly articulate the desired traits and behaviors.

• Early Engagement: Provide high-potential individuals with challenging opportunities and mentorship.

• Immersive Development: Create personalized learning experiences led by senior executives.

• Lead Self, Lead Others: Encourage self-awareness and personal growth as a foundation for leading others.

Empowerment and Feedback: Provide continuous feedback and create opportunities for self-directed learning.

Ultimately, leadership is about creating a ripple effect. It’s about empowering individuals to step up, take ownership, and become the leaders of tomorrow.

Are you nurturing the leaders of tomorrow?

Chart of the week:

Global Investments in Clean Energy Hit New High!

Global low-carbon energy transition investments reached $2.1 trillion in 2024, a record. That’s an 11% y-o-y increase!

Things to know:

• Investment in electrified transport reached $757 billion in 2024, driven by electric vehicles, public charging infrastructure, and fuel cell vehicles.

• Wind, solar, biofuels, biomass, and other renewable energy investments totaled $728 billion. The earth is greening!

• Renewables, energy storage, and electric cars receive $1.93 trillion in investments. Hydrogen, carbon capture, and sustainable shipping saw a 23% investment drop.

• China leads with $818 billion in clean energy, up 20% from 2023.

However, there’s still a long way to go. From 2025 to 2030, the world must invest $5.6 trillion annually to reach global net-zero by 2050. The journey to fully decarbonizing the economy is just beginning!

Sound Bites you must know:

• Cerebras Systems will host DeepSeek R1 in the U.S. – 57x faster than NVIDIA’s GPUs, promising to keep AI data secure and local.

• Researchers successfully used stem-cell-derived heart muscle patches to improve damaged hearts, a step forward in regenerative medicine.

• Galileo launches Agentic Evaluations to fix AI agent errors before they cost you

• Salesforce finds that 93% of IT leaders see value in AI agents but struggle to deliver

• MIT’s bug-sized robot can fly 100x longer than prior prototypes and has potential application in pollinating crops to help save bee populations.

• Tesla will launch an “unsupervised, no one in the car” robotaxi service in Austin, Texas, in June 2025.

Thank you, for reading and supporting our work!

Team,

Binary Circuit

Greenlight

Schedule a free discovery call

Welcome to Binary Circuits’ 6^th weekly edition

This week's must-know developments:

• DeepSeek's step change in AI efficiency: A small Chinese team trained LLM for $5.6 million to challenge OpenAI. $2.19 per million tokens vs. $60 for OpenAI.

• Google improves AI reasoning: Mind Evolution technique improves AI planning and reasoning. It learns and adapts like humans. Report Word.

• Meta's AI removes language barriers: SeamlessM4T translates speech directly between many languages. This could reinvent global communication.

• Generative AI's inevitable business impact: AI that creates content is shifting industries. Businesses should explore its potential to improve and grow.

Let’s dive in!

Share

What you need to know about China’s latest sensation DeepSeek? AI development accelerates and users benefits the most!

DeepSeek, a lesser-known two-year-old firm spun off from a hedge fund, has caused a stir with its exceptional performance without raising any outside funds. Interestingly, DeepSeek is a team of roughly 200 engineers and PhDs educated in China with no foreign exposure.

DeepSeek released its first LLM in 2023 and then became an overnight success in December 2024 with the release of V3 model which excelled on cost and performance. Interestingly, the company selected to introduce the new reasoning models on January 20th, President Trump’s inaugural day. The following day, Project Stargate was announced by president Trump promising $500 billion investments in AI infrastructure.

DeepSeek R1 vs OpenAI/Meta AI - Note that DeepSeek R1 was trained for $5.6 million, compared to hundreds of millions by prominent U.S. corporations.

DeepSeek-R1 outperforms OpenAI's model in 5 out of 11 benchmarks and closely matches it in the remaining six, demonstrating its competitive edge. It is competitive with the Western models in:

• AIME 2024 (mathematical tasks)

• MMLU (general knowledge)

• AlpacaEval 2.0 (question-and-answer performance)

• Chatbot Arena (UC Berkeley-affiliated leaderboard)

Things to know:

• Cost Efficiency and Performance: DeepSeek's AI models, specifically DeepSeek-R1, saved 96% compared to American models while retaining performance.

• Innovative Architecture: Using a Mixture of Experts (MoE) system, DeepSeek activates only relevant parameters to maximize resource efficiency.

• Open Source and Accessibility: DeepSeek has made advanced AI technology more accessible by open-sourcing its models, enabling AI development and adoption.

Why DeepSeek breakthrough is important – Firstly, it questions whether the billions of dollars of investments by the Western companies in the infrastructure are justified. We think, they are because the lower cost of open-source availability will increase adoption, which will drive more data traffic and need for running inference.

Second, the technological advancements, including Multi-Head Latent Attention (MLA) and DeepSeekMoE, engineering techniques to balance processing between GPU and CPU along with reinforcement learning approach opens the way for scaling of models, which was reportedly becoming challenging.

Third, the DeepSeek breakthrough further closes the AI development gap between China and the U.S. In fact, in the four months since OpenAI's reasoning model o1 debuted in September, the first in the U.S., over ten Chinese businesses have rapidly introduced comparable models. These include Alibaba, Zhipu AI, Moonshot AI, and Shanghai AI Lab. A white paper from the China Academy of Information and Communications Technology, a state-affiliated research agency, reported last year that China accounts for 36% of the 1,328 AI LLMs worldwide.

What can the business leaders learn from DeepSeek? Fundamental AI research and “hardcore innovation” at the architectural and algorithmic level. Additionally, fostering a non-hierarchical, researcher-driven culture that encourages creativity and long-term vision.

Google’s Mind Evolution could be a game changer in improving AI reasoning skills

Mind Evolution improves large language models' (LLMs) responses to planning and reasoning tasks, allowing them to "think" more broadly.

Mind Evolution is driven by two main components:

• Search algorithms, which find the optimum reasoning path, and

• Genetic algorithms, which develop and evolve potential solutions using a “fitness function.”

How does Mind Evolution work?

• Initialization: Creates LLM-generated natural language solutions.

• Evaluation and improvement: LLM assesses and enhances solutions.

• Selection: Chooses next-generation parents depending on solution quality.

• Recombination: Crossover and mutation create new solutions.

• Refine: Refines novel solutions using evaluation.

• Iteration: Repeat until optimal solution or predefined iterations elapse.

Bit technical?Consider this: Consider having various possible answers to an issue, like multiple routes to a destination. Google's Mind Evolution is like a competition where these responses are evaluated and improved. While the bad feedback is discarded, the best "survive" and inspire greater ones. The cycle continues, improving replies until the best one appears. Like how nature grows, except ideas are selected and polished. This allows AI to creatively and successfully solve complex problems without knowing the exact steps.

Mind Evolution’s “evaluation” function, unlike traditional evaluators, works with natural language planning tasks and avoids formalizing problems. It also uses an “island” approach to explore diverse solutions, migrating optimal solutions to create new ones.

Why Mind Evolution matters:

Mind Evolution transforms AI from static models to self-improving systems, increasing strength, efficiency, and adaptability over time. It also enables to:

• Think deeper: It helps the AI explore multiple views and modify its replies, yielding more complex findings.

• Reduces human intervention – Reduces manual fine-tuning, enabling autonomous and scalable AI.

• Adapt to new situations: This approach lets the AI "evolve" its solutions and adapt to new knowledge and circumstances, just like in nature.

• Unlock new possibilities: It could help AI perform more complicated activities like planning, design, and creative problem-solving more accurately and efficiently.

Language Barrier – Meta’s AI is making it a thing of the past!

Language has long been an impediment to seamless global communication, but Meta's new AI model, SeamlessM4T, is groundbreaking. This innovative method can directly translate speech from 101 languages into spoken words in 36 target languages, in a much-needed departure from conventional way of converting speech to text, translating, and creating speech again.

Historically, machine translation, particularly for low-resource languages, faces challenges due to limited labeled datasets. SEAMLESSM4T addresses this by utilizing parallel data mining to collect audio-subtitle pairs, enabling training on a diverse range of languages without human annotation.

Facts to consider:

• 443,000 hours of audio with matching text were collected.

• 30,000 aligned speech-text pairs were created without human annotation.

• The model was pre-trained on 4.5 million hours of spoken audio, allowing it to learn general speech patterns.

• SeamlessM4T consists of three AI blocks, handling both text and speech inputs and outputs for efficient translation.

This faster, more accurate, and more efficient technique puts us closer to real-time translation, enabling seamless global communication for organizations, individuals, and industries.

What's impressive is that it's 23% more accurate than the current top models, and it's almost as fast as a human interpreter. It can handle text-to-text, text-to-speech, and speech-to-text translations with ease.

Why does it matter?

This breakthrough has profound implications for businesses operating in an increasingly globalized world. Here's how SeamlessM4T can be a game-changer:

• Expand your reach: Break into new markets and connect with customers worldwide, regardless of their native language.

• Enhance customer experience: Offer instant support and personalized interactions in your customers' preferred language, fostering stronger relationships.

• Boost collaboration: Facilitate seamless communication between teams and partners across different countries, driving efficiency and innovation.

• Unlock new opportunities: Access valuable insights from multilingual data sources, enabling better decision-making and strategic planning.

Business benefits:

• Breaking Language Barriers: SEAMLESSM4T enables seamless communication across languages, cultures, and borders, unlocking new business opportunities globally.

• New Market Frontiers: By translating low-resource languages, businesses can tap into previously inaccessible markets, customers, and partnerships.

• Staying Ahead of the Curve: Early adoption of SEAMLESSM4T can provide a strategic advantage in the global marketplace, setting businesses apart from competitors.

• Streamlined Operations: By automating translation processes, businesses can reduce costs, enhance efficiency, and reallocate resources to drive growth.

Subscribe now

Why generative-AI is a step change and an opportunity to reinvent for business owners

DeepSeek’s media coverage has put AI in everyone’s dinner conversation. Every person is fascinated and is intrigued by the technology and its potential. We believe businesses embracing innovative technologies flourish in the very competitive environment of today.

With its distinct added ability to generate content on top of analytics, Generative AI (GAI) is a step change for reinventing businesses worldwide.

Should you care?

Generative AI opens novel possibilities rather than simply automating them. Imagine a time when your company can achieve unprecedented levels of efficiency, create new solutions quicker than ever, customize consumer experiences on a large scale, and automate repetitive tasks that require thinking. GAI promises realism like this.

Consider these important industry transformations:

Banking: Improved customer service and efficiency, potentially adding $200–$340 billion annually.

GAI creates regulatory reports automatically, therefore saving time and mistakes. It lets banks adjust by offering actionable insights from real-time and past data. AI chatbots quickly answer customer queries about their financial situation, saving customers money and boosting satisfaction.

Retail: effective processes and unique experiences to add roughly $400–$600 billion.

GAI chatbots respond to common questions around the clock and make intelligent decisions about returns and replacements. GAI examines consumer behavior to focus on personalized campaigns in real-time, increasing sales conversion. GAI dynamically forecasts demand and supply to enhance stock control and lower waste.

Life Sciences: accelerating innovation to proactively diagnose and increase longevity might add $60–110 billion yearly.

GAI speeds up medication development, facilitating life-saving treatments. Personalization, by quantifying health, increases patient happiness. GAI reduces time to market and expenses by simplifying clinical trial design and analysis.

Not knowing where to start is a common obstacle in implementing GAI. However well-thought strategy could open new opportunities. Here are some guidelines to begin with:

• First, identify areas where GAI's knowledge could benefit your firm.

• Invest in training or hire AI experts to provide your workers with essential skills.

• Invite multidisciplinary teams to study and apply GAI ideas.

• Ensure your GAI solutions are responsible and follow all rules.

• Accept your constant learning. AI changes constantly. Check for fresh developments and adjust your strategies regularly.

Ignoring GAI's possibilities is no longer an option. Embracing this technology helps companies boost customer satisfaction, enhance efficiency, and gain a competitive edge.

All set to release GAI's potential for your company?

This is just the beginning. Stay tuned for future posts on specific GAI applications and navigating AI challenges implementation.

At GreenLight LLC, we provide modern tech consultation and integration. If you seek honest recommendations and dependable tools for GAI implementation, your search ends here. Schedule a discovery call now to explore how GAI can transform your company!

Chart of the week:

DeepSeek shatters reasoning costs. DeepSeek R-1 costs $2.19 per million tokens. OpenAI? $60.

That's a 96% price drop!

Sound bites you must know:

• TechWar on steroids!! China announces $138 billion investments in a bold AI push to compete globally

• Tencent, meanwhile, decided to casually shrink 3D design from “days of misery” into “10 seconds of magic”

• OpenAI launches Operator, an AI agent that can do tasks on the web

• The biggest Apple Intelligence upgrade of the year is expected to land in April - here's what you can expect in iOS 18.4

Schedule a free discovery call

Welcome to Binary Circuits’ 5^th weekly edition

This week's must-know developments:

• AI Targets Longevity: GPT-4b, a bioengineering AI model aims for longevity through re-engineering Yamanaka proteins.

• Augmented Mentality Emerges: AI evolves from being a tool to offering cognitive “abilities.”

• Stargate Project Launched: A $500 billion U.S. initiative aims to secure AI leadership and boost economic growth.

• Workforce Transformation:AI is set to create millions of jobs by 2030, with skills like creative problem-solving.

Let’s dive in!

Share

AI sets its eyes on extending human lifespan

OpenAI and Retro Biosciences created a protein-generation machine learning model. Scientists have struggled to study protein unfolding. Google DeepMind shocked the world by open-sourcing a database of all proteins for research.

OpenAI’s GPT-4b model is a bioengineering-focused version of GPT-4o created for Retro. The model predicts protein interactions rather than protein structures, unlike DeepMind’s AlphaFold. This protein engineering project initiated last year was supported with $180 million funding.

Retro Biosciences aims to extend human lifespan by 10 years by focusing on Yamanaka factors, a set of proteins that can transform human skin cells into youthful stem cells.

The promise of OpenAI’s GPT-4b

Natural Yamanaka proteins are ineffective at transforming adult cells into stem cells, making these treatments difficult to develop. Retro scientists aim to re-engineer useful Yamanaka proteins using capabilities of Generative-AI.

Retro scientists are targeting two Yamanaka factor proteins, which are crucial to fetal development, using the model. These proteins cause mature adult cells like skin cells to behave like stem cells, according to previous studies.

Preliminary results show promise

These newly modified versions produced 50 times more biomarkers associated with cellular reprogramming in early lab tests. They appear to make stem cells better than natural proteins.

Why does it matter for longevity research?

Reengineered Yamanaka proteins could help in:

• Leading to new treatments for conditions such as blindness and diabetes.

• Creating functional human organs for transplantation and research.

• Generating cells for regenerative medicine and tissue repair.

What’s next?

The two firms will continue research to validate these discoveries and develop new illness treatments. AI is accelerating scientific advancement, which means that the solutions to longevity will become more common.

Tools to ability: A new era begins in 2025

Until now, technology has provided tools that has helped us do more, mostly in physical world. In last two decades, these tools have increased our reach and connectivity with digitization. The undergoing AI revolution is a step change, where the technology will evolve from providing tools to offering cognitive “abilities.”

This new era of technological evolution is deemed as “augmented mentality.”

Wondering what it is?

The fusion of human and artificial intelligence, where AI assistants subtly shape our decisions, interactions, and thoughts. In this context, AI assistants could anticipate our needs, offer real-time counsel, and enhance our mental capabilities.

Future AI assistants will go beyond basic tasks. They will be able to perform tasks like identifying objects, recalling information and drawing actionable insights in real-time, and providing suggestions on diet, travel, business strategies, etc.

AI integration in consumer products, particularly wearable devices, is increasing. Eyeglasses, equipped with sensors, offer a platform for AI-driven context-aware assistance.

Key developments:

• Meta’s big stride: Meta initially made an impact by adding context-aware AI to their Ray-Ban glasses and showcasing their Orion mixed reality prototype with remarkable visuals. Meta is leading the wearable gadget market with ambitions to enhance its Ray-Ban smart glasses by 2025.

• Google’s response: Android XR, an AI-powered operating system, adds context-aware content to augmented reality. The company has partnered with Samsung to release new eyewear and headsets. Google can challenge Meta with its 70% mobile OS market share and growing AI presence with Gemini.

Many forecasts indicate that by 2030, augmented mentality could become integral to individuals and businesses. Several use cases, especially in retail, marketing, education, and advertising will become possible.

For example: This month, Vuzix Corporation, a leading supplier of smart glasses and Augmented Reality (AR), and its partners displayed a variety of smart glass apps operating on Vuzix smart glasses at the Consumer Electronics Show in Las Vegas. The applications included:

• Enhanced warehouse & logistics operations

• Increased remote support and guidance

• Role in conversational AI & assisted learning

Subscribe now

An essential read if you are ready to scale your business. Lessons from ace entrepreneur and investor, Naval Ravikant

Alamanck of Naval Ravikant, a gem for business aspirants, was recommended to me. Naval Ravikant’s almanack is a guide to creating wealth without getting lucky. Every sentence in the book deserves multiple readings. I blogged about the crucial lesson two years ago. I'm thrilled to share this with you all again here.

Some of the key takeaways (by no means are these the only ones), as excerpted from the almanack, are:

• Creating wealth is a skill anyone can learn and apply. Wealth is a set of assets that make money while you sleep. Money is how we transfer time and wealth. Get wealthy by giving society what it doesn’t yet know how to get — at scale.

• You’re not going to get rich renting out your time. In most cases — renting out your time means you’re replaceable. You must own equity — a piece of a business — to gain your financial freedom.

• You don’t get rich by spending your time to save money. You get rich by saving your time to make money.

• Arm yourself with specific knowledge, accountability, and leverage.

  • Specific knowledge is primary for creating wealth. Pursuing your passion and curiosity rather than the latest trend yields specific knowledge. Thus, developing specific knowledge will feel like play to you but labor to others. No specific knowledge can be taught. In this age of unlimited leverage, intellectual curiosity has never been more profitable. If society trains you, it may train someone to replace you. Apprenticeships teach technical or creative skills.

  • Clear accountability is essential. Without accountability, you lack incentives and credibility. High accountability makes you less replaceable and increases equity. Own your business and take chances. Society rewards responsibility, equity, and leverage. Publicly failing under their names usually gives them power. Nowadays, failure isn't a big deal, so people should take greater responsibility.

  • A leveraged person can out-produce a non-leveraged one by a factor of one thousand or ten thousand. Business leverage comes from; 1) people (the oldest form, such as labor, meaning people working for you); 2) capital (money, which can be bought and deployed for outsized returns); and 3) products with no marginal cost of replication (such as software code and media including internet, which are permissionless and source of all new wealth).

• Productize yourself — Productize = leverage + specific knowledge. Yourself = uniqueness + accountability + specific knowledge. Question yourself — “Is this authentic to me? Is it myself that I am projecting?” And then, “Am I productizing it? Am I scaling it? Am I scaling with labor, capital, code, or media?”

• Judgment is the most important skill. Judgment is the exercise of wisdom, which comes from experience. Decisions are crucial because of the leverage of modern technology, large workforces, and capital in the business. Someone who makes decisions right 80 percent of the time instead of 70 percent will be valued and compensated hundreds of times more in the market. You’ll get nonlinear returns if you can be right and rational.

• Get comfortable with frequent, small failures. If you’re willing to bleed a little bit every day, but in exchange, you win big later, you’ll do better. The crowd is after the reverse — frequent, small victories (don’t be the crowd).

• Escape competition through authenticity. No one can compete with you on being you. When you’re competing with people, it’s because you’re copying them. It’s because you’re trying to do the same thing. But every human is different. Don’t copy!

• The secret to peace of mind: You’re competing against yourself — it is a single-player game. Training yourself to be happy is entirely internal. There is no external progress, no external validation.

Future workplaces emphasize human-machine collaboration over job displacement

The labor market is undergoing rapid change as AI continues to develop. The common perception is that AI, especially generative AI with cognitive abilities will replace humans with AI agents. However, like every other technology in the past, we believe, AI will also create more jobs than it replaces in the long term.

The World Economic Forum’s Future of Jobs Report 2025 finds that AI could create 170 million new jobs globally while eliminating 92 million positions, resulting in a net increase of 78 million jobs by 2030.

What does this mean for workforce planning?

Traditional employment strategies that rely primarily on degrees are becoming obsolete.With 50% of employers already intending to reorient their businesses in response to AI, it is evident that adaptation is no longer an option. To remain competitive, it is critical to determine the most valuable skill sets.

• Creative thinking: AI excels at processing data and drawing insights. With generative AI, the newer models have displayed human like abilities to perform complex calculations and visual effects. However, it lacks critical thinking terms of the context of the project. The World Economic Forum forecast that many of new jobs will require creative problem-solving skills.

  • Action for Businesses: Encouraging a culture of innovation is key. In a recent example, 3M encouraged employees to spend time on innovative ideas, which led to Post-it Notes.

• Curiosity and lifelong learning: The ability to learn new skills quickly, known as meta-learning, is crucial in today's rapidly changing job market.

  • Action for Businesses: Companies can help employees with continuing education. Amazon's Career Choice Program pays for high-demand courses for employees.

• Emotional agility: As per Capgemini, organizations that invest in emotional quotient (EQ) training see improved team performance and higher job satisfaction.

  • Action for Businesses: Team collaboration and leadership can improve with EQ training. Capgemini's emotional intelligence coaching program boosts productivity and reduces attrition.

Leave a comment

“Stargate” - A $500 billion AI infrastructure push to ensure U.S. competitiveness

President Donald Trump announced on January 21, 2025 the start of the massive $500 billion joint venture Stargate Project, meant to transform artificial intelligence (AI) infrastructure in the U.S. Under the project, OpenAI, SoftBank, Oracle, and MGX are collaborating to build cutting-edge data centers for AI technology.

Stargate focuses on establishing cutting-edge AI infrastructure to boost innovation, jobs, and the economy.

Why is it so essential?

• AI Race: Stargate represents a bold step toward safeguarding American leadership in the fast developing field of AI.

• Economic Growth: Technology will boost economic growth and create hundreds of thousands of jobs.

• National security: Stargate uses AI for intelligence and defense to gain a strategic advantage.

Implications for technology and businesses:

The influence of Stargate will knock on several spheres:

Accelerated AI Development: The initiative will give researchers and developers hitherto unheard-of tools, hence enabling AI capability breakthroughs.

New Business Opportunities: AI will enable industry-wide new products, services, and business models.

Rising AI Adoption: By companies trying to get a competitive edge, Stargate will probably help to push more acceptance of AI solutions.

As we embark on this huge exertion, corporate leaders must be prepared to play the long game. Successful people understand when to walk away and when to stay the course.

Strategy for Success

• Ensure Data Integrity

• Foster AI Literacy

• Strengthen Cybersecurity

• Attract and Retain Top Talent

• Embed Responsible AI Practices

Chart of the week: AI continues to attract capital!

Investors are going all-in, pouring billions into AI-focused companies. The numbers are staggering:

• 50.8% of the global VC funding flowed to AI businesses in Q4 2024, almost double the share from 2023.

• AI accounts for 25.9% of deals, up from 21.4% in Q4 2023.

• According to early PitchBook data, global VC funding for AI start-ups grew 52% to $131.5bn in 2024, while other start-ups declined 10% to $237bn.

Soundbites you must know:

• State of the AI report finds that over a quarter of the businesses evaluated haver deployed AI while 14% are in beta testing and 21% are building proof of concepts.

• Netflix's new AI model, Go-with-the-Flow, enables you make videos by creating basic paths for objects to follow. Watch a demo of one of its kind AI program.

• Morgan Stanley Research identifies four key megatrends shaping 2025’s investment landscape: global commerce rewiring, longevity’s impact, shifting energy priorities, and AI evolution, especially agentic AI.

• Google quietly released Gemini 2.0 Flash Thinking, an AI model that explains its reasoning process, a crucial step in increasing transparency and building trust.

Thanks for reading!

Cheers,

Team Binary Circuit

This week's must-know developments:

• DOLPHIN AI: automates scientific research, boosting efficiency and innovation.

• Brain Computer Interface Technology: Enhance autonomy for paralyzed individuals and promises new consumer uses.

• Steve Jobs' Business Philosophy: prioritize core values over product features for stronger customer ties.

• Google DeepMind’s FACTS Grounding Platform: Ensures the accuracy of AI responses, tackling LLM hallucinations.

• Energy Transition to Technology: The shift to tech-driven energy generation to gain momentum.

Let’s dive in.

Share

DOLPHIN, an AI tool, is a massive leap in automating scientific research

With artificial intelligence developments, scientific study is about to undergo a paradigm change. Developed by Fudan University and Shanghai AI Lab, DOLPHIN automates labor-intensive and ineffective scientific research activities.

What is DOLPHIN?

DOLPHIN is the first fully integrated system that automates scientific research end-to-end. Unlike existing tools that focus on specific stages—like idea generation or computation—DOLPHIN combines:

• Idea Generation: Proposes novel research ideas from literature ranking.

• Experimental Verification: Conducts experiments with auto-generated code.

• Iterative Feedback: Refines ideas based on results, fostering continuity in improvement.

This closed-loop system mimics the human scientific process but at a faster and more efficient pace.

Arguably, the iterative feedback function is a game changer in automating scientific research. In initial experiments, DOLPHIN’s iterative feedback loop helped improve the output in:

Debugging Success Rates:

• Initial State: DOLPHIN's first iteration had a debugging success rate of 33.3%.

• Improvement: The system improved by 50.0% through learning from error messages, enhancing reliability and efficiency.

Idea Novelty and Relevance:

• Initial State: The system generated 20 ideas with a 40% novelty rate.

• Improvement: DOLPHIN enhanced originality to 95% and maintained costs at $0.184 per idea by re-ranking literature and removing redundancy.

DOLPHIN’s performance is competitive to state-of-the-art human-designed methods:

• 2D Image Classification: Achieved 82.0% accuracy, matching ResNeXt (82.2%).

• 3D Point Classification: Achieved 93.9% accuracy, surpassing PointNet's 91.0% and closely rivaling GPSFormer's 93.8%.

• Sentiment Classification: Increased accuracy by 1.5%, narrowing the gap between BERT-base (91.0%) and BERT-large (93.1%).

Why DOLPHIN matters:

• Accelerating Research Cycles – DOLPHIN automates tasks like debugging and literature review, significantly cutting research cycle time.

• Improving Efficiency and Accuracy – DOLPHIN automatic feedback loop for improvement enhances quality and scientific rigor in each iteration.

• Democratizing Research – The system’s cost-effectiveness ($0.184 per idea) makes advanced research accessible to smaller institutions researchers.

• Advancing AI-Driven Discovery – DOLPHIN advances autonomous research, allowing AI to innovate and explore actively.

DOLPHIN’s success highlights the potential of closed-loop systems in scientific workflows. Future advancements may encompass cross-disciplinary research, refined feedback metrics, and open-ended innovation.

Brain-computer interface technology could soon move from the lab to the market. The technology could create possibilities for novel business models

Imagine using your thoughts to manipulate tools. Through an interface usually connected to your skull, advances in Brain-Computer Interface (BCI) technology allow direct connection between your brain and outside devices. Many businesses are advancing this idea.

Why does it matter

For those with severe paralysis, BCI technology has shown a remarkable ability to increase their autonomy. One such example is Mark, 64, who battles amyotrophic lateral sclerosis (ALS). Synchron's innovative BCI implant lets him just think to run smart home appliances, stream entertainment, and even video chats. This technique uses generative artificial intelligence to enable digital communication, therefore giving a lifeline to people unable of mobility or speech.

A major turning point was reached in China in December 2024 with the first clinical trial of flexible BCI systems for real-time language synthesis. The amazing speed and precision of this technology were demonstrated by a patient with a brain tumor and epilepsy, decoding 71% of Chinese syllables within a week.

Eliminating the need for open-brain surgery, Synchron's BCI uses a minimally invasive "stentrode" implanted close to the motor cortex BCI is a safer choice since its invention considerably lowers surgical hazards and shortens recuperation time.

Immense Possibilities

BCI technology enters consumer wearables and gaming with futuristic experiences. Imagine thought-controlled gaming or smart headphones that analyze brain activity.

Rapidly improving BCIs signal a new age of human-computer interaction and present great possibilities for innovation in fields including consumer electronics and healthcare. Through thinking, non-invasive BCIs will transform mobility for paralyzed people by allowing control of prosthetic limbs and helpful tools, hence improving their quality of life.

Particularly with language decoding as the next development that could revolutionize contact with our surroundings, BCIs show great promise.

Steve Jobs' Timeless Advice: Promote Values Above Product Specifications

“Marketing is about values,” famously stated Apple’s founder Steve Jobs, one of the greatest visionaries the world has seen. This simple tip is beneficial for all companies.

One could get lost in the demand for advertising and data. Jobs believed that organizations should emphasize their core values rather than product features to connect with customers.

Jobs used Nike as a case study. Nike positioned athleticism and greatness at the center of their marketing campaigns, never about the gear. Their ads emphasize brand values over technical specifics.

Likewise, Apple's “Think Different” campaign honored visionaries, rebels, and inventors under this principle. It reminded consumers that Apple was about innovation and progress, not just hardware.

This customer-centric approach is crucial now more than ever. Companies must express their values in today's competitive corporate environment.

Problem: LLM hallucinations/inaccuracy. Solution: Google DeepMind’s FACTS Grounding platform

LLMs often produce inaccurate or “hallucinated” responses, which raises the trust issue for their widespread adoption in critical applications.

Introducing FACTS Grounding

Google DeepMind launched FACTS Grounding, a benchmark evaluation of LLMs on producing accurate and thorough answers from given materials.

FACTS Grounding establishes a new benchmark with diverse input data from 1 to 32,000 tokens (about 20,000 words) across finance, technology, retail, medicine, and law.

Using the provided public examples and approach will also allow you to assess your model's performance.

How FACTS Work: FACTS Using a dataset of 1,719 examples—each comprising a document, a system instruction, and a user request—grounding finds LLMs must go over the document and produce responses that are:

• Accurate: It checks whether the response is free of hallucinations and grounded in the source document.

• Comprehensive: Attend to the entire requests of the user.

For evaluations, FACTS appoints three top LLMs—Gemini 1.5 Pro, GPT-4o, and Claude 3.5 Sonnet as judges to rate the accuracy and applicability of answers.

Google’s Gemini 2.0 Flash leads the FACTS leaderboard with a factuality score of 83.6%. Other top models are Gemini 1.5 Pro, GPT-4o, and Claude 3.5 Haiku, all over 61.7%.

Why This Matters:

• Restore Trust: Correct answers boost brand reputation and user confidence.

• Compliance: Reliable AI lowers legal and financial risks connected to inaccurate or illegal remarks.

• Client Relationship: Informative chatbots boost client loyalty and propel long-term involvement increase.

• Industry-wide Cooperation: The FACTS leaderboard increases openness and competitiveness, therefore encouraging AI research innovation.

Subscribe now

The inevitable rise of energy as technology: Is 2025 the year that accelerates the transition? Have you strategized net green energy in your business model and investments?

The energy transition to technology from commodity sources is currently underway. However, the key question is not if but when the transition accelerates globally. We think that the need for decarbonization to mitigate climate change and, more importantly, the falling costs of technological sources of energy could mark 2025 as the pivotal year in the migration to greener sources.

Energy has long been associated with commodities like oil, coal, and natural gas. These resources are:

• Finite: Earth-extracted, limited quantity.

• Geographically Constrained: Regional deposits create geopolitical linkages and vulnerabilities.

• Subject to Market Volatility: Supply, demand, and geopolitics affect prices.

However, now the new sources are technologically driven and follow different rules:

• Infinite: Solar, wind, and geothermal are almost unlimited.

• Ubiquitous: Not geographically constrained, enabling energy self-sufficiency.

• Technology-driven improvement in efficiency and cost: Innovation drives down cost as more deployment improves technology and economies of scale.

One difference between fossil fuels and alternative energy is improvement potential. Laws of physics naturally limit fossil fuels. Even with improvements, material limitations keep these plants below ideal maximum efficiency. For example, coal plants run at 30–45% efficiency, much below theoretical limits because of physical constraints.

In contrast, alternative energy technologies like solar, wind, and batteries act more like digital technology. Their costs drop dramatically with each production doubling due to learning curves. Solar and lithium-ion batteries demonstrate a learning rate of over 20%, whereas wind power is 15%. Continuous cost reduction comes from material, production, and design innovation.

This shift has profound implications. In his September paper, Azeem Azhar notes that dollars per barrel will be replaced by dollars per megawatt-hour as electricity becomes the new currency. This change will bring in an era of energy abundance whereby practically every nation can become an energy superpower, therefore reducing the power of conventional energy giants.

Chart of the week:

A recent McKinsey report identifies 18 arenas of growth. These areas are characterized by high growth and dynamism and are industries that capture a significant share of economic growth and experience rapid market share shifts.

Identified potential arenas of the future, such as AI and biotechnology, could generate $29 trillion to $48 trillion in revenue, with profits ranging from $2 trillion to $6 trillion. The formation of arenas is driven by business models or technological changes, accelerating investments, and large, growing markets.

Sound bites you must know:

• 2024 was the warmest year on record, surpassing the 1.5C global warming target set by the Paris Agreement. Meanwhile, energy use is rising.

• Outgoing President Joe Biden announces another round of chip technology export restrictions. Will this one work? Maybe not in the long term, similar to the ineffectiveness of the previous restriction.

• OpenAI’s first AI agent, “Tasks,” launched for beta testing. Like a personal assistant, Tasks allows ChatGPT’s Plus, Team, and Pro users to schedule reminders and actions.

• Meet Omi, an $89 AI wearable with always-on listening technology aiming for thought-reading capabilities.

This week's must-know developments:

• Drones: From Hobby to Essential Business Tool – How Will You Use Them?

• Reading: Your Must-Have Skill in the Digital Age

• NVIDIA Pushes AI Boundaries Yet Again

• Generative AI Cost Wars: China’s Deepseek V3 gives a wake-up call to the U.S.

• AI's Rapid Growth: Don't Overlook Google's Gemini 2.0

Let’s dive in!

Share

Drones are ready to transition from a hobby to a vital component in commercial applications. Monitoring assets to package delivery, the uses are just beginning to unfold. How would you plan to use drones?

Drones have evolved from remote-controlled robots to AI-powered machines. Processing data, making decisions, and adapting to their environment offers military and economic possibilities.

The military is investing more in AI-driven drones for surveillance and tough-environment operations. Recent conflicts like the Russia-Ukraine war have increased their importance. Commercial companies like agriculture and energy use drones to monitor crops, examine infrastructure, and improve resource management, cutting costs and increasing efficiency.

Technological advances in drone delivery, including payload capacity, flight duration, and navigation, increased their relevance and applicability. In 2024, this technology emerged as a major feature of supply chain logistics, notably last-mile delivery for e-commerce, healthcare, and humanitarian relief distribution, signifying a move from experimental programs to broad applications.

With successful deployments worldwide, drones are disrupting delivery services. Walmart in the U.S. delivers 20,000 goods to 36 sites using drones, 6,000 deliveries. Meituan's drone manufacturing hub will have 220,000 deliveries by 2023, an average of 12 minutes. Asia's 7-Eleven and Australia's Doordash are also testing 15-minute drone deliveries.

Recent PwC analysis predicts drone growth. Each delivery drone may reduce carbon emissions by 45 tons per year—equivalent to 1,800 trees—and cut delivery times in half, according to the report.

Other key projections include:

• 808 million drone deliveries expected by 2034 from 5 million last year.

• Delivery costs to drop from $6-25 to $2 by 2034.

• The value of goods delivered by drones is to grow 74% annually from $251M in 2024 to $65B in 2034.

Why reading must be a priority in the digital age

Fewer people are reading books and long-form fiction, hence reading habits are dropping. We no longer find great books to hook us. We are thus not ingesting words and ideas that widen our horizons and encourage critical thought and thorough introspection. Rather, driven by convenience, we are skimming feeds, clicking on links, and reducing our attention span.

Some data points to consider:

• The percentage of U.S. adults who read books has dropped by 7% in the last decade.

• In 2022, only 37.6% of the U.S. adults read a novel, down from 45.2% in 2012.

• Americans spend over 8 hours a day consuming digital media.

The problem is much more acute in the education system and children. In 1976, 40% of high-school seniors read at least six books for fun, while 11.5% read none. Those ratios turned around by 2022. Only 17% of the roughly 300 teachers surveyed in EdWeek say they teach full texts; almost a quarter say books are no longer crucial to their courses. Regardless of intelligence or drive, individuals read less and scroll more, as the chart below shows.

The Oxford University Press word of 2024 was “brain rot”— meaning both the “low-quality, low-value content” found online and the intellectual deterioration from its overconsumption.

A recent Financial Times article quoted Maryanne Wolf, author of Reader, Come Home: The Reading Brain In A Digital World, “…while our brains are primed for language acquisition, they are not innately programmed to read; reading is a learnt skill. But brain plasticity works both ways: use it or lose it, and we are increasingly choosing to lose it.

Adam Karaoguz, writer of Renaissance Humans, argues that “Reading is now a superpower.” We can’t agree more. Reading offers immense rewards. Just six minutes of reading can reduce stress levels by two-thirds. Regular readers report better focus, sleep, and mental well-being. Deep reading enhances critical thinking, empathy, and creativity—skills that scrolling can’t develop.

Reclaim your love for reading in 2025 and encourage your children for it.

Creating an environment for reading is essential. There is an old saying, “Read what you like until you love what you read.” Surround yourself with books. Read to your children. Start small and schedule time. Read short blogs, books, or work-related material. Joining a book club could be helpful, too.

Once more, NVIDIA stretches the artificial intelligence horizon

Launching COSMOS and DIGITS at CES 2025, the AI-chips champion signals that artificial intelligence is becoming a general-purpose technology and will shortly be woven into every thread of our lives.

COSMOS introduces artificial intelligence into the physical environment.

NVIDIA Cosmos uses physical artificial intelligence to enable autonomous automobiles and robots to process and comprehend real-world data. Cosmos enables creators of physically accurate simulations and synthetic data generation for training robots and autonomous cars to be empowered.

This development enables intelligent systems to plan, reason, and function autonomously, hence transforming sectors dependent on automation. Jensen Huang argues that we are in a period of "physical artificial intelligence," a significant departure from observation and generative artificial intelligence to actual interaction.

Early adopters pushing the envelope of robotics and autonomous driving include Agility Robotics, Figure AI, Uber, Waabi, and Wayve using Cosmos.

DIGITS lets artificial intelligence supercomputers run on your desk.

Designed for academics and developers, NVIDIA's Project DIGITS is “the world's smallest AI supercomputer.” May will see the availability of the price at $3,000 powered by the NVIDIA Grace Blackwell Superchip. It allows the execution of huge language models with up to 200 billion parameters independent of cloud infrastructure. This invention democratizes access to sophisticated artificial intelligence capabilities, therefore encouraging broad acceptance in several fields.

NVIDIA also is working with automakers to expedite the creation of driverless cars. With NVIDIA's Orin chips and software, Toyota intends to improve advanced driver assistance systems in the next car generations. In autonomous trucking solutions, Aurora is using NVIDIA's AI systems to increase transportation and logistics efficiency.

The race to reduce the cost of generative AI training and inference. China’s Deepseek V3 gives a wake-up call to the U.S. One thing is clear: whoever wins, the businesses will have plenty of low-cost options for AI integration

Many businesses still consider the high training cost of AI models as their complexity increases and the per-token expense for inferencing a major obstacle to the wide adoption of generative AI. China seems to be leveraging software engineering to bring down costs better than the U.S. While the U.S. is considered a major innovation hub for technological breakthroughs, China takes pride in optimization and makes it scalable by bringing down costs and building applications.

Kai-Fu Lee, an AI expert, former CEO of Google China, and CEO of 01.AI, noted in a conference that their LLM ranks sixth in terms of performance but was trained with just $3 million compared to $80-$100 million for GPT-4. More interestingly, 01.AI reduced the inference costs to 10 cents per million tokens, one-thirtieth of a typical comparable model.

Now, another complex Chinese model, DeepSeek V3, with 685 billion parameters, is outperforming its Western counterparts in performance and cost. DeepSeek v3 is remarkably cost-efficient, with a training cost of approximately $5.6 million and inference rates as low as $0.14–$0.28 per million tokens.

The new model excels, surpassing leading open models like Meta’s Llama-3.1 and Alibaba’s Qwen, and closely competing with closed-source models like OpenAI’s GPT-4o and Anthropic’s Claude 3.5. Its unmatched performance in Chinese and math-focused benchmarks scored 90.2 on Math-500, while Qwen scored 80.

The speed of AI advancement is remarkable. Why should you also consider Google's overlooked Gemini Deep Research with Gemini 2.0?

Google's Gemini Deep Research tool, part of Gemini 2.0, changes internet research. AI analyzes over 50 web pages to generate insights and citation-based reports. Unlike other AI assistants, it creates a customized research plan to achieve desired results.

Unique advantages

Deep Research offers real-time Google Search and relevant information. It chops up complex searches to refine results and yield insights that would take days to obtain manually.

Users can export results to Google Docs for study and collaboration. Free for the first month, $20 per month, it works with Open AI Pro ($200) and Google products for personalization.

It's promising compared to other LLMs

Compared to other AI research tools, early users saw impressive results:

• Income Stream Surfers: Wrote a 3,000-word article from one search without editing.

• Teacher's Tech: Showed how targeted queries yield additional sources and richer insights.

• Olivia Sarikas: Excellently evaluated tweet performance using Deep Research to understand viral content.

Why is this important?

Gemini Deep Research addresses the demand for fast, reliable, automated research within a data-driven environment. This capability allows consumers to make swift, data-informed decisions, enhancing industry competitiveness and innovation.

Deep research can significantly shape our approach across various fields. For example, students can quickly grasp and learn challenging subjects, while professionals can utilize it to create reports, research markets, and analyze trends. Furthermore, content creators, authors, and academics can employ it to gather facts and write effectively.

Chart of the week:

Microsoft is the bell weather of AI development with massive investments. Is the company overly optimistic, or is it able to see the future better than its counterparts?

Microsoft was the largest buyer of NVIDIA chips in 2024 and is also the biggest spender on AI infrastructure. The big-tech giant has announced $80 billion in capex for 2025.

This week's market-moving developments:

• Why NVIDIA's acquisition of Run:ai is a massive opportunity for businesses of all sizes

• Anthropic's Model Context Protocol (MCP) is a game changer for AI. Enterprise AI integration made easy

• Post-quantum cryptography is an emerging field for innovation. What should you know?

• AI agents are poised to become mainstream in 2025

Let’s dive in!

Share

Why NVIDIA's acquisition of Run:ai is a massive opportunity for businesses of all sizes

Run.ai, an Israel-based company, specializes in Kubernetes-based orchestration, therefore enabling dynamic GPU allocation across workloads by optimizing GPU resource use. This increases the scalability of artificial intelligence systems by improving efficiency in AI model training and application. NVIDIA received regulatory approval on December 30 for this $700 million acquisition.

Run.ai’s software complements NVIDIA's GPUs and AI systems to attract additional customers. The acquisition strengthens NVIDIA's market dominance by reaching mid-sized companies and startups without large-scale AI infrastructure.

The collaboration will aid the democratization of AI across business sizes and geographies. Here’s why.

• Simplifes AI infrastructure management: It lowers the demand for costly in-house IT personnel, therefore allowing smaller companies and academics access to AI resources.

• Accessibility to AI infrastructure: Accessibility expands to companies of all sizes because of its availability on-site, in the cloud, or in hybrid configurations.

• Open-source platform increases innovation and customization: Reports imply Run: AI aims to open-source sections of its platform to foster innovation and make tools available to more developers.

• Cost Efficiency: Run.aI's dynamic resource pooling reduces GPU idle time, making AI operations cheaper.

• Broader Developer Access: It helps developers build and grow AI apps quickly without GPU management experience.

  Share Binary Circuit

  ---

Anthropic's Model Context Protocol (MCP) is a game changer for AI. Enterprise AI integration made easy

Last week, Anthropic published MCP, an open-source tool, MCP that could transform AI system data connectivity.

What is MCP?

Consider designing an AI assistant. It must be connected to client databases, financial information, cloud storage, etc. to be useful. Developing custom connectors for every data source takes time and effort.

Unlike building separate connections, MCP can easily combine all data sources. Using a common protocol, MCP links artificial intelligence systems to any data source, therefore facilitating this. Consider it as an AI universal adaptor.

Why does it matter?

• Efficiency: MCP removes custom connectors, saving developers time and money

• Scalability: It simplifies connecting AI systems to many data sources, enabling more complex and powerful applications

• Interoperability: MCP standardizes AI system-data connections, improving AI platform interoperability

• Innovation: By simplifying data integration, MCP expands AI applications and fosters creativity, enabling previously unthinkable applications

Business Impact:

• Efficiency: Deploy AI solutions across different data sources without bespoke coding for faster implementation

• Reduced Technical Debt: Standardized protocol reduces maintenance

• Future-proofing: As AI systems evolve, the protocol remains consistent

MCP could transform AI system development and deployment. We can anticipate:

• Smarter AI assistants as more data means more knowledge and ability

• Easier integration will help organizations integrate AI into their workflows

• MCP could help unify and interoperate AI ecosystems

Steve Jobs' Timeless Advice: Promote Values Above Product Specifications

Timeless Steve Jobs Marketing Advice: Values, Not Specs

"Marketing is about values," famously stated Apple founder Steve Jobs. This simple tip is beneficial for all companies. One could get lost in the demand of advertising and data. Jobs believed that organizations should emphasize their core values rather than product features to connect with customers.

Jobs used Nike as a case study. Nike positioned athleticism and greatness at the center of their marketing campaigns, never about the gear. Their ads emphasize brand values over technical specifics.

Likewise, Apple's "Think Different" campaign honored visionaries, rebels, and inventors under this principle. It reminded consumers that Apple was about innovation and progress, not just hardware.

This customer-centric approach is crucial now more than ever. Companies must express their values in today's competitive corporate environment.

Post-quantum cryptography is an emerging field for innovation. What should you know?

Leading-edge quantum computing advances are moving the technology closer to practical applications. Google, IBM, and Microsoft reached milestones in 2024. Google's Willow device decreases scaling qubit noise and errors, while IBM's Heron processor executes complex algorithms with unprecedented accuracy. These advances push quantum computing toward general deployment, increasing the need to safeguard communications and data post-quantum.

Quantum computers threaten the conventional RSA and ECC encryption. For example, Shor's algorithm is a quantum program that can crack various cryptographic protocols, which secure financial, healthcare, and government data. One rising issue is the “harvest now, decrypt later” strategy, where attackers acquire encrypted data now and decrypt it later as quantum capabilities develop.

Post-quantum cryptography (PQC) is a potent solution to resist quantum attacks. The U.S.-based National Institute of Standards and Technology (NIST) released its first three PQC standards in August 2024. These standards significantly protect digital infrastructure from future threats. NIST approved 14 more quantum-resistant encryption methods in October 2024. Governments and businesses are prioritizing PQC investments to defend national security and finance.

NIST has established post-quantum cryptography milestones. RSA-2048 and ECC-256 algorithms will be discontinued by the end of this decade, and by 2035, they will be completely banned. Quantum-resistant system migration is urgent for enterprises, given this timetable.

Established companies are already adopting PQC. Amazon Web Services announced its migration strategy to safeguard its services using PQC techniques, and SEALSQ launched SEALQUANTUM, a post-quantum lab. These efforts emphasize the need for enterprises to build migration methods and quantum-resistant algorithms.

AI agents will become mainstream in 2025

AI agents are poised to completely change the way that tasks are completed. Startups and big internet companies are investing billions, and 2025 will be their breakthrough year.

The AI technologies of today are, at most, copilots, helping users but needing supervision. That might not be the case soon. AI agents, according to experts, will function autonomously, analyzing data, making judgments, and acting without human direction.

The promise is enormous. AI agents might conduct data analysis, sales, and customer support, among other chores. Salesforce and Microsoft offer tools to build custom AI agents. SalesForce CEO Mark Benioff stated that he aims to deploy one billion agents within a year. OpenAI intends to launch “Operator,” its AI agent, this year.

Investments are on fire. Startups focused on artificial intelligence agents saw investment rise 80% in 2024. Deloitte's Global 2025 Predictions Report notes that 25% of generative AI users will deploy AI agent pilots this year, and the second half of the year may see full agent adoption in some workflows.

The workplace could shift radically once these AI agents become as successful and autonomous as intended. One thing is certain: AI agents will rule conversations in 2025 regardless of success or failure.

Chart of the week:

Generational AI became fully integrated with digital twins in 2024, setting the stage for an explosion of innovative use cases in 2025, when these dynamic duos will foresee failures, optimize designs, and even create new products.

This week's market-moving developments:

• NVIDIA's $249 Edge Processing Revolution: 70T ops/second transforms enterprise AI deployment economics

• OpenAI O3's Quantum Leap: Human-surpassing reasoning capabilities demand rapid strategy shifts

• Intelligent Enterprise Acceleration: Digital twins + GenAI convergence creates unprecedented optimization potential

• Humanoid Robot Economics: Cost curves suggest imminent market disruption; $24T TAM by 2035

Let’s dive in!

Share

Breakthrough of the week: NVIDIA launches Jetson Orin Nano Super

The processor performs 70 trillion operations per second, runs LLMs, consumes 25 watts, and sells for ONLY $249 (watch video). Imagine the possibilities for your business. What would you do with this processor? We can help you build use cases.

The Jetson Orin Nano Super is a versatile tool for designing LLM chatbots, visual AI agents, and AI-based robotics. This development will take GAI to the edge devices.

Imagine humanoid robots with embedded intelligence and IoT sensors that can run the LLMs on-premise without requiring private data to be transferred to the cloud. The applications are limited to our imagination.

Open AI’s O3 Breakthrough Shows Technology is Accelerating Again

OpenAI model o3 outperforms humans in math and programming with its reasoning. It scored an impressive 88% on the advanced reasoning ARC-AGI benchmark. A big improvement from 5% earlier this year and 32% in September. Codeforces placed O3 175th, meaning it can outperform all humans in coding.

Why Does It Matter?
Most notable is the shorter development cycle. O3 launched months after O1, while previous AI models required 18-24 months.

Unlike conventional models that take months of training, o3 enhances inference performance. This means discoveries might happen in weeks, not years. This rapid technology advancement will require organizations to reassess innovation timeframes.

Impact on Businesses
Many complex subjects can be developed faster after the o3 breakthrough.

• Scientific Research: Accelerate protein folding, particle physics, and cosmic research

• Engineering: Quick prototyping and problem-solving

• Math: Access to previously inaccessible theoretical domains

• Software Development: Enterprise-grade code automation

An Enterprise Competitive Playbook: Implement AI reasoning tools in R&D pipelines quickly. In the coming years, organizations will have to restructure tech teams around AI, and AI-first R&D will become mainstream.

Intelligent Enterprises to Go Mainstream

An intelligent enterprise uses advanced technologies like AI and digital twins along with real data to simulate scenarios, design products, and predict cons and demand. Embedded intelligence helps optimize, improve, and come up with new ideas.

Why Does It Matter?
By 2030, enterprises will have "data omnipresence." These organizations will use Agentic AI, tiny language models, and quantum computing to automate with human oversight.

The Digital Twin’s Role
Businesses can mimic and optimize operations before implementation using digital twins. Digital twins reduce production errors, development time, and costs. By 2027, McKinsey expects the digital-twin market to reach $73.5 billion, expanding 60% yearly.

The Convergence Effect
Adding genAI to digital twins adds value:

• GenAI compresses and organizes heterogeneous digital twin datasets for better predictions and anomaly detection

• GenAI’s natural language interface allows smoother communication with digital twins, boosting usability

• AI-generated code accelerates digital-twin development, lowering costs and complexity

Strategy for Success
Digital transformation demands planning and leadership. Key strategies are shown by Sikorsky Aircraft:

• Virtualize every part of product lifecycles

• Test and improve items with digital twins before manufacturing

• Empower cross-functional teams to innovate and break silos

• Track progress to show ROI and maintain momentum

Humanoid Robots Adoption is at an Inflection Point.

Elon Musk and Brett Adcock, CEO of Figure AI, predict as many as 10 billion humanoid robots by 2040. That’s more than the global human population. MetaTrend Robotics Report provides insights that declining costs and improving technology embedded into robots are making them work more closely with humans.

Characteristics of humanoid robots: From basic housework to massive industrial processes, these intelligent, human-like machines can do a range of tasks. With 1,000 Optimus devices placed at their factories this year, companies like Tesla are already setting the trend.

Dr. Peter Diamandis, CEO of XPrize, says that the question isn't whether this revolution is coming. It's whether you'll be prepared to capitalize on it.

Why now?

Generative artificial intelligence, high-torque actuators, and more computing capabilities taken together have accelerated the creation of humanoid robots. These robots are whole functioning systems able to do difficult jobs.

Additionally, costs to manufacture them are expected to come down multifold. From $250,000 now, to $150,000 in a year. This is a 40% decline compared to the typical 15%-20% decline witnessed in recent years. Tesla ambitiously expects its humanoid robot costs to decline to $20,000 over the next few years.

Important Takeaways:

• By 2035, estimates of market growth range from $38 billion (Goldman Sachs) to $24 trillion (Ark Invest)

• Robots could affect 40% of workers and 75% of professions, helping mitigate labor shortages in manufacturing, elder care, and agriculture. By some estimates, in the U.S. alone, there are roughly 8 million job openings for the work no one wants to do

• Companies like Figure AI are drawing large capital; their values run at $2.6 billion

Business ramifications:

• Robots can run around-the-clock, therefore cutting expenses and speeding manufacturing cycles

• Aging populations and declining birth rates create demand for robotic solutions, particularly in logistics and healthcare, which help to address labor shortages

• Affordable robots could enable generally available services, hence reducing costs and increasing production

  ---

Chart of the week:

Batteries resume cost declines. Overcapacity to some extent, but the trend is unstoppable over the medium term

Bloomberg NEF analysis shows that the lithium-ion battery packs cost $115 per kilowatt-hour, down 20% from 2023.

This drop is attributed to cell production overcapacity, economies of scale, decreasing metal and component prices, and the use of cheaper lithium-iron-phosphate (LFP) batteries.

BNEF reports 3.1 terawatt-hours of fully commissioned battery-cell production capacity worldwide, more than 2.5 times the 2024 lithium-ion battery demand. Would this gap narrow down as mitigating climate change becomes even more urgent?

This significant price drop predicts that electric car prices could achieve parity with internal combustion engine vehicles by 2026, with average battery pricing below $100/kWh.

Subscribe now

A bit about GreenLight—the engine behind Binary Circuit. GreenLight finds possibilities, analyzes challenges, and develops plans to fit and grow firms to stay ahead. Digital transformation, corporate intelligence, and AI are our specialties. Helping businesses succeed is our goal.

What should you expect?

• Analysis of major technical advancements and their implications.

• Stories about organizations adapting—or failing—in transition.

• How to future-proof your business.

Use this guide to thrive in rapid technological shifts by discovering your edge and capitalizing on it.

Thank you for being here. Let’s explore what’s next.

Share Binary Circuit