“Give me a lever long enough and a fulcrum on which to place it, and I shall move the world.” — Archimedes Archimedes would be all-in on AI 😄

Building AI products is nothing like shipping traditional software. A powerful strategy is to create AI systems that are on the edge of feasibility. What's barely possible today could become effective in six months when a new model is shipped. The reason it's different is that

I've jotted down some insights on Bitmapist, our open-source cohort analytics tool that's already saved us millions and guides many decisions across nearly every Doist team. Not many people know about it, but you should check it out! We also open-sourced bitmapist-server, which

We’re rapidly approaching the era of general artificial intelligence. A question on my mind is how we’ll stay relevant (from company, product and individual standpoint). I think it comes down to three key questions: - Will smarter or cheaper AI directly benefit our company? -

Todoist Ramble is coming to mobile. I'm highly biased and very excited, because this feels like a next-gen AI interface that isn't just a chatbox. Imagine a world where AI is no longer a selling point, but woven into everything. We'll have much smarter tools that handle all the

No wonder stuff is getting done on the Base app 🚀✅😄!

Keep an eye on this. It will blow your mind.

PageRank and Transformers look different on the surface. Under the hood, they run on the same foundation: math. PageRank ranked the web by solving an eigenvector problem on a giant graph. Transformers (the core behind most LLMs) understand language by turning words into vectors

how can engineers (or anyone) get better at design? i get this question a lot. as someone who went from cs to design, here's a real path: start with systems thinking engs already get this. design is just systems for humans & our senses, instead of machines. read “thinking in

Todoist is becoming liquid soon 💦! In general, we have some pretty incredible things coming up!

Some major 🚩 flags in this line of thinking: - Code produced is a vanity metric. You can produce a ton of garbage code, and LLMs will do this. - The real challenge isn't just building things, but maintaining them. If you rely solely on LLM-generated code, you'll accumulate

A demo of Quick Add using Apple's on-device Foundation Model to suggest a project automatically. The future will be wild for app development! 🔥 🎩-tip to .Enric Enrich and Witold

Showcasing the latest Todoist Ramble improvements (you can already try them out if you're an experimentalist and paid). We are also planning to make it possible to specify projects and labels via rambling. The team is really on fire 🔥

Core takeaway from the GPT-5 launch: The French will never give up the dream of everyone speaking French 😂

I've had preview access to GPT-5 for a couple of weeks, so I have a lot to say about it. Here's my first post, focusing just on core characteristics, pricing (it's VERY competitively priced) and interesting details from the GPT-5 system card simonwillison.net/2025/Aug/7/gpt…

I have been using GPT-5 today in Cursor and doing various research/tasks in ChatGPT… and it’s pretty underwhelming. It feels like a Llama 4 moment where they were more focused on benchmarks than usefulness. It’s especially bad at coding; it spends too much time thinking over

Unpopular opinion: Apple’s Liquid Glass design system is incredible, and this will be a huge design milestone. The best part isn’t just the material or transparency, but all the nice transitions and animations. There is a lot of focus on content. I’ve been using it since the

Got a new action on my lock screen! 🗣️☑️