I’m a Claude code power user. I just gave codex my first spin and it fixed a bug in an hour that I had been working on in CC for days. Will be using this more. Going to try the CLI next. Am I missing out on anything else?
2026 prediction: Chat won’t be the main way we use AI.
The biggest shift over the next 5 years? Video and 3D becoming the primary interface to powerful models. Text stays relevant, but it won’t be the growth story.
Seeing amazing 1-shot apps built with Gemini-3, while I burned all my tokens debugging @linear MCP integration with @antigravity. Works fine in terminal but won’t connect through the agent. Back at it in 5 hours when tokens refresh.
@evanlapointe Reminds me of the Amazon Operating Principle “Leaders Are Right, a Lot”. Nikita has excellent judgment which means velocity might seem like the main component for success when really it’s table stakes and strong judgment is the differentiator.
I’m getting back into 3D in a big way. Excited for what is to come. Getting close to wrapping up my first right of passage. Thanks @andrewpprice for all you do for the blender community!
I have to give credit where it’s due. The @X team has done a really fantastic job on the For You algo. I’ve really noticed a quality bump even on niche topics I’m really interested in. @nikitabier props to you and the team.
Product sense in action! I haven’t been able to commit to an AI browser because I keep missing all the little things that made Arc magical. Love what the @diabrowser team is doing.
Another big hit from Arc was "Auto-Join" for Google Calendar meetings. Well, looks like it's coming to Dia!
Join Zoom and Google Meet video calls in one click... coming (back) to a browser near you!
This is counter-intuitive especially as orgs get larger, but ultimately is the right answer.
The downsides typically get painted over in large orgs because of their existing competitive advantages and the structured appearance of winning matters more (promos, resourcing, etc.)
Goals on product teams only make things worse.
You either hired right, or you didn’t. That is the only variable controlling the quality, quantity, and pace of product output.
Layering a goal makes every great product person worse, unless they totally ignore it. And it doesn’t make any bad product person good.
Great read on how bubbles can create long term benefits. The primary example referenced were the cable companies that went bankrupt during the dotcom era that laid the backbone of the internet.
The Benefits of Bubbles
We are in an AI Bubble: the big question is if this bubble will worth it for the physical infrastructure and coordinated innovation that result?
https://t.co/bmHFQQrclV