A BEST CASE SCENARIO FOR AI? The Doomer narratives were wrong. Predicated on a “rapid take-off” to AGI, they predicted that the leading AI model would use its intelligence to self-improve, leaving others in the dust, and quickly achieving a godlike superintelligence. Instead, we are seeing the opposite:  — the leading models are clustering around similar performance benchmarks; — model companies continue to leapfrog each other with their latest versions (which shouldn’t be possible if one achieves rapid take-off); — models are developing areas of competitive advantage, becoming increasingly specialized in personality, modes, coding and math as opposed to one model becoming all-knowing.  None of this is to gainsay the progress. We are seeing strong improvement in quality, usability, and price/performance across the top model companies. This is the stuff of great engineering and should be celebrated. It’s just not the stuff of apocalyptic pronouncements. Oppenheimer has left the building.  The AI race is highly dynamic so this could change. But right now the current situation is Goldilocks: — We have 5 major American companies vigorously competing on frontier models. This brings out the best in everyone and helps America win the AI race.  As @BalajiS has written: “We have many models from many factions that have all converged on similar capabilities, rather than a huge lead between the best model and the rest. So we should expect a balance of power between various human/AI fusions rather than a single dominant AGI that will turn us all into paperclips/pillars of salt.” — So far, we have avoided a monopolistic outcome that vests all power and control in a single entity. In my view, the most likely dystopian outcome with AI is a marriage of corporate and state power similar to what we saw exposed in the Twitter Files, where “Trust & Safety” gets weaponized into government censorship and control. At least when you have multiple strong private sector players, that gets harder. By contrast, winner-take-all dynamics are more likely to produce Orwellian outcomes. — There is likely to be a major role for open source. These models excel at providing 80-90% of the capability at 10-20% of the cost. This tradeoff will be highly attractive to customers who value customization, control, and cost over frontier capabilities. China has gone all-in on open source, so it would be good to see more American companies competing in this area, as OpenAI just did. (Meta also deserves credit.) — There is likely to be a division of labor between generalized foundation models and specific verticalized applications. Instead of a single superintelligence capturing all the value, we are likely to see numerous agentic applications solving “last mile” problems. This is great news for the startup ecosystem.  — There is also an increasingly clear division of labor between humans and AI. Despite all the wondrous progress, AI models are still at zero in terms of setting their own objective function. Models need context, they must be heavily prompted, the output must be verified, and this process must be repeated iteratively to achieve meaningful business value. This is why Balaji has said that AI is not end-to-end but middle-to-middle. This means that apocalyptic predictions of job loss are as overhyped as AGI itself. Instead, the truism that “you’re not going to lose your job to AI but to someone who uses AI better than you” is holding up well.  In summary, the latest releases of AI models show that model capabilities are more decentralized than many predicted. While there is no guarantee that this continues — there is always the potential for the market to accrete to a small number of players once the investment super-cycle ends — the current state of vigorous competition is healthy. It propels innovation forward, helps America win the AI race, and avoids centralized control. This is good news — that the Doomers did not expect.

🧵 The Origins of “Magic Internet Money” Where did the Bitcoin Wizard come from? And why is $MIM reviving the most iconic meme in crypto history? 👇 1/15

As models get more powerful, i find myself focusing more effort on context engineering, which is the task of bringing the right information (in the right format) to the LLM. Context engineering is hard because it is pervasive. You need to engineer every layer of the stack to capture and make context available. Send too little context, and the LLM won’t know what to do. Too much, and you’re out of tokens or the LLM gets lost. Good context engineering caches well. Bad context engineering is both slow and expensive.

SOUND ON. You’re hearing the first howl of a dire wolf in over 10,000 years. Meet Romulus and Remus—the world’s first de-extinct animals, born on October 1, 2024. The dire wolf has been extinct for over 10,000 years. These two wolves were brought back from extinction using genetic edits derived from a complete dire wolf genome, meticulously reconstructed by Colossal from ancient DNA found in fossils dating back 11,500 and 72,000 years. This moment marks not only a milestone for us as a company but also a leap forward for science, conservation, and humanity. From the beginning, our goal has been clear: “To revolutionize history and be the first company to use CRISPR technology successfully in the de-extinction of previously lost species.” By achieving this, we continue to push forward our broader mission on—accepting humanity’s duty to restore Earth to a healthier state. But this isn’t just our moment—it’s one for science, our planet, and humankind. All of which we love and are passionate about. Now, close your eyes and listen to that howl once more. Think about what this means for all of us.

MCP is going viral. AI agents can now talk to real tools & apps and actually get stuff done. This changes everything. 10 wild examples: