Claude writes about 90% of Claude Code's code. From June to September, merged PRs more than doubled while the share reverted fell by more than 40%. We wrote about how we changed code review to get there, and how to adopt it: claude.dev/blog/solving-code…

I'd expand the moats left to: - Proprietary data or content: data nobody else has, or unique content you can create - Network effects / community: users make it more valuable for each other - Audience / distribution: you can get attention on demand - Physical component: hardware, locations, real-world operations - Capital: owning assets and investing (esp into the AI companies as they are and will suck up large parts of the economy) - Regulatory: licenses and legal barriers

cloud agent enjoyers watching people still walk around with their laptops cracked open

I've been thinking a lot about this. Buckle up. Back in the 90s and early 2000s, most devs worked internally at companies, building custom, on prem, internal software. Fast forward to the 2010s, and SAAS companies explode in popularity, replacing the need for so many internal devs, because now you build the CRM once, and resell it to thousands of companies. You might think this would lead to fewer devs... But as we know, the SAAS age was incredible for dev careers. Now, software is even cheaper. Cheaper even than reusable saas. We're gonna build more software internally again. Not all of it will be internal, saas isn't totally dead, but weird one offs that might have been kinda jankily solved with a kinda-related SAAS, can now be solved with bespoke internal software. The cost of that software IS NOT zero. People saying that don't understand the concept of zero. But it is cheaper. I predict small teams of internal devs building incredible amounts of highly custom internal software will be a thing. At nearly every company of every type, as long as it's more than a ~30 person org.

Talked to a finance guy today who said his firm has not only totally stopped buying software in favor of building it all in-house, they've actually hired multiple SWEs to do the buildouts, even though their salaries cost more than the subscription fees they were paying. Having tools custom-built for their specific business is just that valuable. So they're not spending less money or saving time, versus the pre-AI world, but they're getting a much higher return on the money and time they are spending. Much to ponder

I've been using AI agents to run a lot of my life lately. First Instinct, now Muse. Hotels, restaurants, stuff I buy - all over chat. Used to work well, but things are changing right now as agents are getting blocked. Both are free. Meta pays for Muse. VCs pay for most of the rest. That won't last. Someone ends up paying, and the obvious candidate is the hotel or store on the other side. That's the problem. If hotels pay my agent per booking, is it booking me the best room - or the one that pays the most? Same if I ask for a stroller: the best one, or the brand paying the biggest cut? With Google I see a page of options and the ads are labeled. With an agent I see one answer. My bet is it goes the other way. Agents pool demand. 5,000 people want a beach week in Cabo (Mexico) in November. 50,000 parents need a stroller this month. The agent takes that to the sellers and they compete for it. The agent gets paid for bringing buyers, not for steering them. That's the one I'd want working for me.

We announced an agreement to acquire a nationwide low-band spectrum license portfolio that will pave the way for @Starlink to become a major mobile carrier in the US. With this new spectrum and our Gen2 constellation, Starlink Mobile can ensure Americans have access to high-speed mobile broadband no matter where they are → spacex.com/updates#starlink-…

The developer who started the project of cloning Photoshop and other Adobe projects with AI is now going after Microsoft Office. WordCraft is here. Like other Craft projects, this one is in Rust too and this one is in Alpha too. Start is a good thing. But will we ever see the finished product?

🚨Breaking: OpenAI and Anthropic executives are reportedly privately preparing for catastrophic AI events According to Axios, they're planning for the fallout of a major AI cyberattack that could potentially take down banking, internet or power Absolutely insane that we're already at this point, many AI industry insiders apparently believe a major incident will happen within the next 6 to 12 months

I don't think I've seen a topic as wildly, incapacitatingly, intellectually deranging as Agentic Commerce. Note that there are no arguments here; it is taken for granted that a) "people" (who?) will want agents to buy "stuff" (what stuff?) for them and b) "they" (who?) won't want to use "some Amazon-supplied agent to do it" (why not? And how do we interpret Amazon's in incremental annualized sales in 2025 from Rufus / Alexa for Shopping then?). This is not disciplined, logical analysis. This is faith; religious fervor. That some new technological capability becomes the dominant consumer form factor is asserted as inevitable without support. But we know what retail platforms' incentives are: to directly manage consumer relationships so as to induce broader monetization, advertising revenue, and deeper integration (see: Amazon and Walmart commentary on the value of advertising to their businesses, Amazon's deterministic identity spine reaching 90% of US households, Walmart's acquisition of Vizio and vibe, etc.). If all of this is stripped away: what actual opportunity exists to compete with Amazon? In other words, if the consumer relationship is competed away, what's the value of merely operating the agent <-> retailer plumbing? Sounds like a commodity with substantial operating costs if the retail platform is still expected to manage fulfillment and logistics, customer service / returns, etc. Or will the agent do this?

Amazon banning agents is the first opportunity I've seen since Amazon was founded for a startup to create an Amazon competitor. People will want agents to buy stuff for them. It will be one of the main use cases. And they won't want to use some Amazon-supplied agent to do it.

I just tried the photoshop clone, and it's completely unusable. The thing people don't understand about software development is the bottle neck wasn't ever the implementation, but rather the care and consideration behind every choice.

it's unlike Anthropic to NOT have a response to GrokBot / Dots what's going on over there 🤔