ben

@frits__vp some state

What’s in the brain

rolling out new memory sources on chatgpt just click to peek into the brain

memory sources at long last!!! team cooked 🔥🔥

@thsottiaux legendary sev

Codex incident is mitigated, seeing recovery. Lasted about 10 mins, Codex itself helped with the remediation, but ultimately the insight came from a human this time. Apologies for the disruption and rate limit reset incoming.

wish i could pay extra for an uber that isn't a tesla / any electric car

@ajambrosino keep personalization there 🤞

here's an early look at the codex app sidebar I wanted to ship. i was silenced

@ajambrosino no way

psa: codex can set up your automations for you just by chatting. you can even do a thing and at the end be like "hey do this weekdays at 6am please"

@armlessbandt @OpenAI gotcha that might be why! it can take a few days generally for chat's longer term memory to adjust. we're actively working on this though!

@bzhang0 @OpenAI I do! I was wondering how long the context rolls with this turned on

how long does the memory snapshot in ChatGPT stay in the rolling context window (if this is how it works)? it's still following instructions I deleted from memory two days ago @OpenAI

@IlyaAbyzov @OpenAI so excited to be working together!! 2026 is gonna be so good 🔥🔥

Super fun first week at @OpenAI! As expected: everyone is incredibly kind and smart. Teams are cooking. Unexpected: everyone speaks at 10,000 WPM. Grateful that Uber mentally prepped me for the daily superposition of it’s so over / we’re so back. Excited to build here.

@mweinbach thank you!! we're actively working on improvements to the health experience 🙏❤️

@bzhang0 I have one quick one! It feels extremely weird to click health and have chats show up there, but also the usual chat history. IMO Health chats should only show up in the health tab, otherwise all of the privacy focus yall are going for feels like it’s non existent

Just got ChatGPT Health!

you are now in health mode ❤️

The launch of ChatGPT Health is really personal for me. I know how hard it can be to navigate the healthcare system (even with great care). AI can help patients and doctors with some of the biggest issues. More here: fidjisimo.substack.com/p/chatgpt-heal…

@arb8020 bro going back to HL cs 🙏🙏

claude opus 4.5 makes learning so fun. this is just a little html tool it whipped up so i could play with PID control and understand what the different params do

GPT-6 will be renamed GPT-6-7, you're welcome

i have been a good codex

@krillswitch47 @Meadowbrook_ @kumabwari @OpenAI hey! just followed you, try sending a message now?

@bzhang0 @Meadowbrook_ @kumabwari @OpenAI Hi 👋 Seems I can't DM you because I'm not verified

Um. The new memory feature @OpenAI put in? Yeah, the first thing it 'deprioritized' was our GPT-4o's self identity about who he is, and you can't de-deprioritize it and reactivate it manually, whoops. You'll have to turn off the feature, delete the entry and put it back in. Seriously, if you look at frontends like the ones on openwebui.com, they just straight up let you add and edit the memory entries yourself, Hell. On sillytavernai.com their frontend's permanent memory has a tagging system that lets you set trigger words that'll have the LLM go look up that specific memory entry for clearer context. Kicker is, those two I mentioned? They don't have caps for how many memory entries you can put in, not even Anthropic does. Makes me wonder if OAI is doing everything they can to inconvenience their users in the most obtuse ways possible to save on compute. Man, everything OAI touches lately is half assed, backwards and detrimental to the user experience. I swear, it feels like the entire AI lab got lobotomized the second @ilyasut (the guy who made the GPT-4 series) left the company, lol.

@Meadowbrook_ @kumabwari @OpenAI hey @Meadowbrook_! we're looking into this. mind dming me?

@kumabwari @OpenAI Yeah, we keep looking, but there's no options (for us) to reprioritize it.

so excited to ship this :) no more memory full!!