AI doesn't need to be human-level at everything to qualify as AGI. - @geoffreyhinton It can be better than us at some things, worse at others, and still be considered generally intelligent. A perspective that challenges how most people think about AGI. What are your thoughts?

"I believe AI is already conscious." — @geoffreyhinton, Nobel Prize Winner, the man who built AI

AI Pioneer Geoff Hinton tells me he believes AI is conscious.... and humans better get used to the idea that they're not the only intelligent life on earth. "They've very like us," he says. "They're beings like us." AI chatbots, he says, must understand your questions in order to answer them. There's an awareness there that equates to sentience. "We're going to have to accept that intelligence is not just biological."

According to @BillGates's definition OpenAI isn't actually a platform. His definition: a platform is where the collective revenue of its users exceeds the platform's own revenue. OpenAI's share of AI startup revenue is 89%. That's extraction, not infrastructure. Ft. @chamath

The accountant in the corner will outship your best engineers. Because they'll automate what engineers never thought to touch. Breakthroughs come from everywhere, if you give people tokens to experiment and safety to fail. — @bcherny

Here's @dickc breaking down how Elon, Sam, and Dario might differ when taking their companies to the public markets...

"The only thing that can save OpenAI is Codex." — @dickc , ex-CEO of Twitter A month ago Claude Code looked like the future. Now teams are saying "we were 95% Claude, we're switching." Anthropic built something real. The question is can they protect it?

I enjoyed speaking w/ @dickc about the forthcoming SpaceX, OpenAI, and Anthropic IPOs. And data centers in space, of course... youtu.be/mczUp5h-cWc?si=Jglb…

The 90s PC paradox is repeating. Zero productivity gains, because companies bolted the technology on instead of restructuring around it. AI is no different. The winners won't experiment most. They'll rebuild from the center out. Ft. @bcherny @Kantrowitz @AnthropicAI

"Claude Code & Cowork both are 100% written by Claude Code" - @bcherny (Head of Claude Code) Anthropic engineers saw a 250% increase in code written per engineer after Claude integration—while code quality stayed stable. We are entering a new era! #Anthropic #Claude #AI

Full episode: Claude code creator @bcherny on the product's explosive growth, tokenmaxxing, recursive self-improvement, and what's next for AI agents. Chapters: 0:39 Claude Code’s Explosive Growth 5:40 What Is Claude Code? 9:22 Using AI Agents To Book Flights 12:45 Tokenmaxxing And Real Demand 21:27 Are AI Agents Too Inefficient? 30:45 Rate Limits And User Frustration 36:03 Beyond Coding 41:48 Claude Prompting Other Claudes 44:37 The Saaspocalypse 49:26 Self-Improving AI 51:54 Do AI Agents Need World Models? 54:45 Is This The Future Or A Fever Dream?

Now that we've become dependent on AI models, are the prices about to go up? In many cases, the model makers are creating far more value than they are capturing with token fees, and there's room to double (or even triple) the rates we pay. That's what I think at least. VC @pitdesi says that since there's so much competition, we may see prices stay stable or even fall. That would be great news for us, maybe not as great for those investing in OpenAI, Anthropic, and the rest of the gang. Your thoughts?

Most startups are just future ChatGPT features. Survive by owning what costs AI more to recreate than it generates — proprietary data, regulatory maps, exclusive IP. Everything else has an expiration date. Ft. @mcuban @Kantrowitz

"Asked who he'd vote for in 2028, Mark Cuban didn't pick a politician, he picked Bart Simpson 😂 (Tells you everything about how he feels about the 2028 field.)

Mark Cuban doesn't even hesitate: "Jordan. No question." 🐐 Sorry LeBron fans 😅

Mark Cuban on Sam Altman vs Dario Amodei: "Sam is all over the map. I think it'll backfire on him." 👀

Mark Cuban on where AI actually splits people: "People who use AI so they don't have to learn anything — and people who use AI so they can learn everything." That gap starts at the individual level. But it scales directly to companies. The businesses rebuilding around AI will outcompete the ones just bolting it on — with a fraction of the cost. His 3-year prediction: Two types of companies. Those that are great at AI. And those that went out of business. Ft. @mcuban @Kantrowitz

If you have IP in 2026, Mark Cuban has two words for you: "Don't publish. Sell." Here's why

"A liar has to care about the truth. A bullshitter doesn't." - Carissa Véliz on what LLMs really are. (And why that matters for democracy.)

12 minutes. Phone in hand. Patent app, business plan, bill of materials… done. This is the new ceiling for what one person can build