Pradeep Goel

What’s interesting is that software developer employment has continued to grow alongside these systems rather than contract. That suggests AI may currently be functioning as capability expansion. Lower software development costs can increase demand for software creation itself, opening new categories of products and workflows that were previously too expensive or time-intensive to build. AI diffusion is becoming an economic distribution story.

A report from @Microsoft Global AI Diffusion shows global AI usage rising from 16.3% to 17.8% of the world’s working-age population in just one quarter, with 26 economies now seeing AI adoption rates above 30%. What stands out is not just the pace of adoption, but where the acceleration is happening. The report highlights strong growth across Asia, particularly in South Korea, Thailand, and Japan, driven in part by improving multilingual AI capabilities.

@anndylian The future probably belongs to people who can build and adapt at the same time.

The future belongs to builders.

@johnennis The human brain really is extraordinary. Especially when you realize it can learn continuously, adapt to completely new environments, and operate on the power of basically a light bulb.

One thing that working on AI has given me is an immense respect for the human brain We all carry around a relatively small mass of organic matter in our heads that uses a tiny fraction of the energy of our most complicated computers, yet can perform apparently arbitrary tasks with immense efficiency And our memory works very well too relative to AI memory systems On top of all that, it also controls our bodies and can move us through arbitrary and novel physical movements The whole thing is completely amazing

@DeryaTR_ I think calling AI just another tool understates the shift, but calling it a species creates confusion too. We’re dealing with a new form of intelligence infrastructure, and we still don’t have the right language for it.

What most people still don't understand about AI is that they think of it like any other transformative technology, such as electricity, computers, the internet etc. AI of course needed all that but it’s not just another tech, it’s a non-biological species at or above our level.

@rohanpaul_ai AI is starting to do to knowledge work what industrial automation did to manufacturing. Not replacing every worker overnight, but radically changing the economics of scale and efficiency.

Satya Nadella reveals how Microsoft is applying the concept of "Lean for knowledge work" internally with AI. The internal ROI on AI investment and leveraging the cost reduction effect of AI. Borrowing from Toyota's manufacturing efficiency principles and applying them to white-collar operations powered by AI. e.g. Microsoft spends approximately $4 billion per year on customer support operations. By deploying AI agents for front-end deflection (resolving issues before they reach human agents) and real-time reasoning assistance for support staff, they are dramatically reducing costs in areas like Xbox and Azure support. --- From "Bg2 Pod" YT channel ( link in comment)

@haider1 Human intuition struggles with exponential change. By the time something feels sudden, it has usually been improving gradually in the background.

Demis Hassabis says the hardest part isn't the technology — it's that human brains think linearly, while AI is an exponential technology But the one advantage we have is that everyone can use it now "that familiarity is what prepares us for the scale of what's coming"

@brian_armstrong The interesting part is that most of these are no longer technology problems. They are coordination, regulatory, and trust problems.

Major areas where the financial system still needs an update: 1. Tokenization of real-world assets - Real estate, stocks, bonds, funds, etc. onchain for instant settlement, fractional ownership & massive distribution. 2. 24/7 Global trading - Pooled global liquidity, every asset, every person, with great leverage and capital efficiency. 3. Next-gen payments - Near-instant, low-cost global transfers using stablecoins, including for Agentic payments. 4. AI-powered risk, credit, compliance, and advice - Better decisions, less fraud, and broader access to capital. Everyone gets access to a great financial advisor. 5. Innovation friendly regulation - Move from one-size-fits-all to risk-based rules that encourage innovation and competition instead of stifling it. 6. Expanded access - Open protocols that reduce middlemen and self-custodial wallets to expand access to everyone with a smartphone. 7. Capital formation - Low cost and turnkey for anyone to raise money for a good idea, increasing the number of startups. 8. Sound money - A refuge from inflation, when discipline is lost in fiat money. Jobs not done until we get these working for all. Will require lots of tech innovation and policy work to get there.

@mark_k This is why reasoning benchmarks are less important than verification loops now

Anything that is verifiable and can be trained with RL is now becoming solvable for AI. Math proofs. Coding. Algorithm discovery. Games. Formal verification. Some parts of science. Once you can check the answer cheaply, AI can search the space at absurd scale. The rest remains elusive.

@daniel_mac8 I think the mistake is assuming AGI would arrive as a finished state instead of an ongoing transition. Most people are still waiting for a dramatic moment, while the systems are already becoming genuinely useful across a wide range of cognitive tasks.

AGI is here. Marc Andreessen is right. Most people didn’t notice because AGI was never going to arrive in a discrete moment. It wasn't go to be televised like the Super Bowl. There is no discrete moment when a child becomes an adult either. We create legal thresholds because society needs them to function, but nature does not move that way. A 17-year-old and an 18-year-old are not different species. AGI is like that, because AGI is a product of nature. People always talked about it as if one day a machine would cross a bright line and everyone would agree: "there it is!" In hindsight, it's obvious that was never going to happen. If you've been working with these systems since 2015, like me, the continuum is obvious. They did not go from useless to magical overnight. They kept getting better and better. It was a slow progression. There were step changes, but the overall progress was smooth. And now they are even superhuman in some areas. > They can build complex software. > They can find software vulnerabilities humans missed for decade. > They can disprove mathematical conjectures that got human attention for 80 years. You can argue about the definition. It's fun. But that argument is becoming disconnected from reality. AGI is here. There, I said it. That's the hill I'll die on. I’m not sure the world is ready. But reality doesn't care if you're ready or not. I have two kids. I never felt ready to become a father. Then it happened, and I adjusted because that is what intelligent creatures do. AGI will require the same kind of adjustment, except the thing arriving is even MORE alien than a baby. And let me tell you man, babies are PRETTY ALIEN!

@rohanpaul_ai I think people confuse intelligence with biology because humans have been the only reference point for so long. The question is not whether machines can become intelligent, but whether we are ready to redefine intelligence itself.

Ex-Google executive Mo Gawdat defines Intelligence. "If we accept that intelligence itself is not a physical property, then it does not matter whether that intelligence is produced on carbon-based computer structures like humans, silicon-based computer structures like today’s hardware running AI, or quantum-based computer structures in the future. Intelligence is produced within machines when we stop imposing our own intelligence on them." --- From 'The Diary Of A CEO and Mo Gawdat' YT channel (link in comment)

The question of data ownership in healthcare is both technical and philosophical. Who truly owns your health data? The hospital that collected it? The company that analyzed it? Or you? Decentralized Science (DeSci) offers a compelling answer: putting individuals at the center of their own data ecosystem. This is not only about privacy, it's about empowerment. Imagine a future where patients control who accesses their health data, are compensated for its use, and can contribute to research on their own terms. This is the inevitable evolution of healthcare.

In healthcare, AI is improving existing systems and expanding access to expertise that was once limited by geography, cost, and infrastructure. But the breakthrough will not come from models alone, AI becomes exponentially more valuable when paired with healthcare systems that understand local realities, resource constraints, and patient behavior.

The strategic value is in controlling the infrastructure layers those models depend on. This is also why hyperscalers are investing heavily in custom silicon like TPUs. Owning the chip layer improves cost efficiency, optimizes workloads for AI-specific operations, and reduces dependence on external semiconductor supply chains. AI competition has evolved into a race for compute ownership, energy access, and deployment scale.

Google and @blackstone are launching a new AI cloud infrastructure venture backed by an initial $5 billion commitment to bring 500 megawatts of data centre capacity online by 2027. The venture will combine large-scale data centre infrastructure with access to Google’s TPUs through a compute-as-a-service model. Cloud competition was largely about software ecosystems and developer adoption, now AI is changing that dynamic. The bottleneck is increasingly becoming physical infrastructure including, compute capacity, power availability, cooling systems, and semiconductor access.