Applied AI products supporting new model releases isn't what is used to be. It's no longer a question of evals, it's a question of agility and, in some cases, grit. Opus 4.5 exposed an entirely new concept called "effort", a parameter that involved optimization and use case

Opus 4.5 is one of the most capable models we've seen yet. Try it out in @NotionHQ

This is the third time I go to X to ask @NotionHQ when a new model is going to be supported, only to find they already announced it. 🙌

Engineering our systems to work and take advantage of different model functionalities is an endless task. Gemini 3 Pro is an impressive model for your hardest enterprise work questions and I think our users will love what it can do. Try it out on @NotionHQ and tell us what you

Boba heard @NotionHQ loves dogfooding and vibecoded his way onto payroll.

notion ai feels way faster to use than any AI chat product. AND it has all of your context from notion/slack built in. 🤯

The trio is Quinn, Modi, and my dynamic coffee-DietCoke duo. I’m just here for the vibes.

It’s nice when smart people say nice things about what you’re working on. Notion is the only S-tier (and one of the few on the list) that wasn’t born as an AI company. It was early to rebuilding what we now know as agents that can do everyday work you previously did manually

Sat down with @OpenAI to talk about building Notion agent and our collaborations leading up to GPT-5.

We are trying to reimagine what search journeys look like in the age of agents and connected workspaces. @NotionHQ is hiring multiple roles on our search team: EM, UX, Ranking, Indexing platform. Come join us.

Agents are all figuring out the web and having their AOL moment, relearning search for agent-first environments. In two years, this will feel like table stakes. In the meantime, I sat down with @paraga/@p0 and @DhruvBatra_ /@yutori_ai on how we get there.

“there are no women in SF” wdym? look at all these women on the TL that just moved to sf and want to meet you and do fall activities!!

Hiii. Just moved to SF and passionate about coding and startups. I love to travel and do pattern recognition. Hit me up if you want to get coffee.

Tell me you’re at a dog park in SF without telling me you’re at a dog park in SF

SF-core

accidentally made it into forbes japan with @sarahmsachs -- sniped candidly sitting on the couch in a fun article about the @NotionHQ office design :)

3/ Another slept on topic is the latency / reasoning trade off. There’s no single “fast enough.” For simple queries, every millisecond matters. For reasoning tasks, users will happily let agents run in the background. The real challenge is designing expectations, not just

2/ We dug into evals in this convo. They’re rarely framed right. Most teams treat evals as one job, but you need two: - Regression + observability evals to keep quality stable and understand what users need - Frontier evals to measure new capabilities and recognize real

1/ Too many “agents” are just prompt workflows in disguise. Reasoning models need orchestration, memory, and modularity. Not chains of prompts and few shots. We rebuilt our AI system from the ground up to make that possible. Sometimes the only way forward is to start over. It

🧵Sat down with @VentureBeat about the engineering approach behind Notion 3.0. Here's what I think not enough companies are willing to do, and why @NotionHQ never looked back