Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.

WebMCP is available for early preview → goo.gle/4rML2O9 WebMCP aims to provide a standard way for exposing structured tools, ensuring AI agents can perform actions on your side with increased speed, reliability, and precision.

I'm too right wing for the left and I'm too left wing for the right. I'm too into humanities for those in tech and I'm too into tech for those in the humanities. What I'm learning is that failing to polarize is itself quite polarizing.

Teaching an experimental class for MBAs on “vibefounding,” the students have four days to come up and launch a company. More on this eventually, but quick observations: 1) I have taught entrepreneurship for over a decade. Everything they are doing in four days would have taken a semester in previous years, if it could have done it at all. Quality is also far better. 2) Give people tools and training and they can do amazing things. We are using a combination of Claude Code, Gemini, and ChatGPT. The non-coders are all building working products. But also everyone is doing weeks of high quality work on financials, research, pricing, positioning, marketing in hours. All the tools are weird to use, even with some training, but they are figuring it out. 3) People with experience in an industry or skill have a huge advantage as they can build solutions that have built-in markets & which solve known hard problems that seemed impossible. (Always been true, but the barriers have fallen to actually doing stuff) 4) The hardest thing to get across is that AI doesn’t just do work for you, it also does new kinds of work. The most successful efforts often take advantage of the fact that the AI itself is very smart. How do you bring its analytical, creative, and empathetic abilities to bear on a problem? What do you do with access to a very smart intelligence on demand? I wish I had more frameworks to clearly teach. So many assumptions about how to launch a business have clearly changed. You don’t need to go through the same discovery process if you build a dozen ideas at the same time & get AI feedback. Many, many new possibilities, and the students really see how big a deal this is.

Sharing an interesting recent conversation on AI's impact on the economy. AI has been compared to various historical precedents: electricity, industrial revolution, etc., I think the strongest analogy is that of AI as a new computing paradigm (Software 2.0) because both are fundamentally about the automation of digital information processing. If you were to forecast the impact of computing on the job market in ~1980s, the most predictive feature of a task/job you'd look at is to what extent the algorithm of it is fixed, i.e. are you just mechanically transforming information according to rote, easy to specify rules (e.g. typing, bookkeeping, human calculators, etc.)? Back then, this was the class of programs that the computing capability of that era allowed us to write (by hand, manually). With AI now, we are able to write new programs that we could never hope to write by hand before. We do it by specifying objectives (e.g. classification accuracy, reward functions), and we search the program space via gradient descent to find neural networks that work well against that objective. This is my Software 2.0 blog post from a while ago. In this new programming paradigm then, the new most predictive feature to look at is verifiability. If a task/job is verifiable, then it is optimizable directly or via reinforcement learning, and a neural net can be trained to work extremely well. It's about to what extent an AI can "practice" something. The environment has to be resettable (you can start a new attempt), efficient (a lot attempts can be made), and rewardable (there is some automated process to reward any specific attempt that was made). The more a task/job is verifiable, the more amenable it is to automation in the new programming paradigm. If it is not verifiable, it has to fall out from neural net magic of generalization fingers crossed, or via weaker means like imitation. This is what's driving the "jagged" frontier of progress in LLMs. Tasks that are verifiable progress rapidly, including possibly beyond the ability of top experts (e.g. math, code, amount of time spent watching videos, anything that looks like puzzles with correct answers), while many others lag by comparison (creative, strategic, tasks that combine real-world knowledge, state, context and common sense). Software 1.0 easily automates what you can specify. Software 2.0 easily automates what you can verify.