If you've ever seen someone tweet some cool shader and thought "I don't really even know what a shader is and at this point I'm too afraid to ask" - I've written something just for you. makingsoftware.com/chapters/shade…

oh you’re using claude code? everyone’s using open code. just kidding we’re all on amp code. we’re using cline, we’re using roo code. we just forked our own version of roo. were using kilo code. we were on coderabbit but their ceo yelled at us so now we’re using qorbit. apple

So, Manus AI sold for $1-2B to Meta. A few of my takeaways to people in silicon valley and not: 1. Manus treated distribution as a first-class expense, spending heavily on creators to win attention early 2. That spend worked because creators showed the product in use, not

New episode w Adam Marblestone on what the brain's secret sauce is: how do we learn so much from so little? Also, the answer to Ilya’s question: how does the genome encode desires for high level concepts that are only seen during lifetime? Turns out, they’re deeply connected

2026 is the GREATEST time to build a startup in 30 years I’m 36. I’ve sold 3 startups, helped build companies that raised billions, and backed teams from seed to unicorn. 20 MEGA shifts that make this the BEST time to build in a GENERATION: 1. Hardware got smart. Download

This guy literally dropped the best visual guide to LLMs you’ll ever see

2026 is going to be the year of Agentic Interfaces Ever since I started using Claude Code with Obsidian, i've been thinking about the best way to set this up on the web (in a sandbox). Basically... what would be the best agent interface? Designed to be used by you and Claude

R.I.P few-shot prompting. Meta AI researchers discovered a technique that makes LLMs 94% more accurate without any examples. It's called "Chain-of-Verification" (CoVe) and it completely destroys everything we thought we knew about prompting. Here's the breakthrough (and why

x.com/i/article/2009…

The biggest opportunities for AI startups today We surveyed my readers about how they're using AI today, and more importantly, how they want to be using AI. For PMs, the biggest opportunity is research. User research shows the largest demand gap of any task. Only 4.7% say it’s

Great resource if you’re planning to get started with agents in 2026

I’ve been building Dexter for 2 months now. It’s like Claude Code, but for finance. What Dexter can do: • find undervalued stocks • analyze them in detail • build investment thesis All of the code is open source. Bonus: Dexter can also run on local LLMs.

Repeating your prompt can make LLMs significantly more accurate. Google just showed a trivial change that wins 47 of 70 tests. No extra tokens. No added latency. Zero losses reported. 𝗣𝗿𝗼𝗺𝗽𝘁 𝗿𝗲𝗽𝗲𝘁𝗶𝘁𝗶𝗼𝗻 𝗶𝗺𝗽𝗿𝗼𝘃𝗲𝘀 𝗮𝗰𝗰𝘂𝗿𝗮𝗰𝘆 The method is simple.

Great piece that highlights the next frontier is long running agents. A current constraint to that future is multi-agent coordination, including sharing context.

Today we're launching our first data partnership with Similarweb Now, you can: 👉 Access 12 months of web traffic history 👉 Benchmark competitors instantly 👉 Analyze marketing channels and traffic sources 👉Get regional traffic breakdowns All powered by Similarweb's

Dream Engineering, a field pioneered by Media Lab Professor Pattie Maes (@fluidinterfaces) and alumni Adam Haar Horowitz and Judith Amores, has been named as one of the 25 most powerful ideas of the 21st century (so far) by BBC Science Focus Magazine. media.mit.edu/articles/the-2…

Apps aren’t disappearing — they’re dissolving The gap between intent and usability is collapsing Building an app as a fixed UI stops making sense AI won’t be products you open. It’ll reshape itself around what you’re trying to do UI isn’t the product. Capability is.

Looking to talk to builders working on long-horizon AI systems. If you’re dealing with: memory over time context drift / forgetting agents that run for hours or days I’d love to compare notes and pressure-test assumptions. Not hiring (yet). Just learning from people in the

I am often asked how has AI changed the way that I work. Two ways: I do not ever type emails. Wispr Flow perfectly transcribes them. Structures them. Spelling of names is perfect. Could not live without Wispr both on iOS and desktop. Every single 20VC show we now run the same

A K shaped future makes sense since Team B loses to Team A 10 times out of 10, and now coding AGI is an accelerant