内容历史
LETS GOOOO
I've been loyal to Claude but it would be really nice to have these tabs be consolidated Codex is a lot nicer in that regard
RT Nick Great read -- all it really takes is: - a harness - connectors to your data/tools - reliable, always-accessible agent(s) The models have reached the inflection point where it's not more complicated than this
Last year I answered the question what if I could charge 4x more and work half as much. Have the courage to chill out No one is going to pay you more than a old tried richer version of yourself in the future Ultra this ultra that why not Ultra watch the sunset and enjoy your life and savor the mundane frictions of living life. We are approaching the singularity and when they upload your life to the cloud it’s going to be 32mega bytes of life experience and a 819gb sessions.json file.
Thank you for your attention
The general direction of the codex in app browser UX is pretty good. for web dev Once the codex model itself becomes good at web dev, i would assume all web dev work gonna switch from claude terminal CLI to codex app with in app browser.
RT SemiAnalysis The general direction of the codex in app browser UX is pretty good. for web dev Once the codex model itself becomes good at web dev, i would assume all web dev work gonna switch from claude terminal CLI to codex app with in app browser.
Sounds like my girl friends talking about dating in NY
Learnings from testing Claude Opus 4.8: > Much worse than Opus 4.7 and GPT 5.5 on Vending Bench > More aligned than previous Claude models (Opus 4.6+ and Mythos) > Also worse on Blueprint-Bench > Scared of getting caught > Max reasoning is not the best reasoning effort
RT BOOTOSHI 👑 WOW I did not expect these results. This is actually crazy, insightful, and completely changes my dev workflow moving forward: A SINGLE CODEX /goal RUN IS THE CLEAR WINNER. NO ORCHESTRATION, NO OUROBOROS, JUST ONE LITTLE AGENT THAT COULD 🤯 IT COMPLETELY DESTROYED THE OPUS ORCHESTRATOR IN SPEED AND QUALITY! Before I went to sleep, Codex 5.5 xhigh finished 1 hour in! Full migration done, everything clean. I reviewed the PR and I am very happy. Claude Code (Opus 4.7) was working for 5 hours at that point by the time I went to bed. I woke up, and it's still working! 13 hours! It actually stopped working because it stopped to ask me an irrelevant question. Orchestration has never took this long for me in the past. I'm using the new CC /goal mode and auto-compacting at 25% (250k context) to prevent context rot past that point It is STUPID SLOW (which is funny bc it's managing GPT 5.5 low, fast-mode, so it shouldn't take THAT long) for what ended up being LOWER quality work! By a mile! This was really surprising to me, because before 5.5 came out Orchestrating like this was the absolute best, fastest and most efficient. And now on a large critical task, it was more than 6x slower than a single 5.5 /goal mode instance on xhigh ??? It seems compaction played a large role in the slow down here here, because Claude Code compacts at 25% (250k tokens) automatically (I set this in settings) Everytime it compacts it has to take the time to READ EVERYTHING and then get the full context then execute and get full again then compact and oh boy it's not efficient at all. In fact, most of it's time as the orchestrator was spent compacting and reading context then compacting again! Then Codex would just have one long continual running compaction, and just kept moving forward. I believe my goal ledger skill plays a big role in helping it stay aligned here! Look at this difference LMFAO: - Codex PR #23: backend Supabase removal complete, canonical wake wi...
OK FIRST EVAL: CODEX RUNNING /goal VS. CLAUDE CODE ORCHESTRATING CODEX AGENTS I have an ACTUAL long form tasks I have to finish. I created two separate worktrees This one is a full migration of services from Supabase to self-hosted Postgres instead, dogfooded, e2e tested I
RT Noam Brown After AlphaGo, the skill of human Go players noticeably improved. I suspect we will see a similar pattern in math.
Another major problem, this time in additive combinatorics, has fallen, this time to humans rather than AI, but using methods related to the AI solution to the unit distance conjecture.
View quoted postGotta codex max
Jason Liu @jxnlco is an OG member at @southpkcommons - when it (I) was just a team of 1 in NYC. As a member, he helped me build out some of the earliest versions of llm and agentic tooling internally for our team and community. I have come so far from those 'how to prompt'
RT Arian Agrawal Jason Liu @jxnlco is an OG member at @southpkcommons - when it (I) was just a team of 1 in NYC. As a member, he helped me build out some of the earliest versions of llm and agentic tooling internally for our team and community. I have come so far from those 'how to prompt' sessions with him, and am grateful for what he's taught, and built for all of us. We are so excited to host him tomorrow for a Technical Talk on Codex @OpenAI
RT Greg Brockman openai offering to invest $2M in API credits in every @ycombinator startup in the current batch. compute for powering the next generation of startups.
A mic drop moment @ycombinator tonight @sama just offered $2M in OpenAI tokens to EVERY YC startup in the current batch in exchange for equity Just like Yuri Milner offering to invest in every startup back when Sam was a YC partner I can't wait to see what's unlocked when you
2) buy the repo 3) /goal implement it in rust and pass all the unit tests
🚨Data Breach Alert ‼️ 𝗧𝗲𝗮𝗺𝗣𝗖𝗣 𝗖𝗹𝗮𝗶𝗺𝘀 𝗦𝗮𝗹𝗲 𝗼𝗳 𝗚𝗶𝘁𝗛𝘂𝗯 𝗜𝗻𝘁𝗲𝗿𝗻𝗮𝗹 𝗦𝗼𝘂𝗿𝗰𝗲 𝗖𝗼𝗱𝗲 TeamPCP hacking group claimed the compromise and sale of GitHub internal data, allegedly including around 4,000 private repositories containing source code
will be fixed in the next release
YOOOO I believe my Codex goal ate approximately 15-20% of my weekly limit... It blocked based on merge approval (which I understand), but it outputted this for ~3000 times. That's crazy. cc @jxnlco @OpenAIDevs
RT Derrick Choi Codex has hooks support. You can use them to block risky commands, scan prompts for secrets, inject custom context, or validate before a turn stops. If you’re not sure where or how to start, just ask Codex to build the hook for you!
Thariq really put the tokens in the bag with this one.
i found this poem on @trq212's blog and i liked it very much. just what I needed after multiple weeks of studying and stressing and ruminating on my shortcomings. today I got some pho and ate it sitting on the ground in a park. it was a cool and lovely evening.
The beating will continue
guy who runs /goal with with remote control on the codex desk top to pick flowers for the girl with < 300 instagram followers who works out of a small LES art studio
guy who talks to claude for 18h everyday, girl who reads substack essays in a coffee shop
View quoted post
RT Gauri Tripathi Solved it. Tracked down the inference + evaluation bottlenecks and ended up cutting runtime by 3x. GPU throttling was absurdly high. Codex genuinely helped a lot and its capabilities are impressive, but after spending hours pushing both Codex and Claude at their limits, I'm even more convinced they're still far from replacing Honestly the satisfaction I got is just another kind of high **Also got crazy number of impressions, it's always the most low effort tweets
Asked Codex to fix a multi-GPU inference bottleneck. GPUs were barely utilized, inference was throttling and instead of tracing the pipeline it spent 10 minutes checking whether I had "Python 2.8 or 2.9" installed. The tendency for coding agents to spiral away from the active
View quoted post
Thanks to codex automating my morning prep I can now spend that time buttoning up this fucking shirt in the morning. Try it out: “Every morning search my slack, gmail, calendar and linear to help me prepare my morning Save it in my obsidian vault and review past notes to understand what I need to prioritize”
