AI (and how we are learning to apply it) is evolving quickly – sometimes in ways we expect and sometimes in ways that completely surprise us. These advances are not only changing our products but also the way we work and solve problems too.
Here are some of the most interesting AI-related finds I caught up on this month. I look forward to hearing what you think!
Anthropic put their latest model, Claude 3.7 Sonnet, to the test to better understand Claude’s extended thinking capabilities by having it play Pokémon! This was incredibly cool – Claude was able to advance through the game and even successfully battle Pokémon Gym Leaders and win badges. While it couldn’t beat the game on its own, I think it demonstrates the promise of some of the latest models…
Satya joins the Dwarkesh Podcast to discuss Microsoft’s vision for AGI and quantum computing. He shares the latest quantum breakthroughs and their impact on AI-powered apps and business transformation. I was particularly interested is his take on cognitive labor: “Don’t conflate knowledge worker with knowledge work.”
As AI agents take on many of our routine tasks, traditional SaaS products will become less visible backend utilities, with agents handling interactions instead of people. This shift might commoditize many SaaS tools, changing the functions of UI, brand, and user experience. This thought-provoking deep dive is quite simply a good read.
Read: What Happens to SaaS in a World with Computer-Using Agents?
It is not only our customers who are changing – but we at Microsoft and the wider tech industry are changing too. At some companies, PMs and domain experts (rather than engineers) are increasingly responsible for the prompt engineering that defines how AI apps behave. And engineers are spending less time coding and more time shaping user needs.