error501
69 posts
error501
@losterror501
codex has become my tiktok
Washington, DC Katılım Mart 2026
158 Takip Edilen39 Takipçiler
Code and instructions to reproduce here, GPT mini can set it up even
github.com/0xSero/codex-s…
0xSero@0xSero
DeepSeek-V4-Pro & Kimi-K2.6 running in Codex app. Cheapest way to taste the frontier. Works w local models and they can do all the computer use stuff (:
English
@ChiragAsarpota wym by opensource + pro/pricing? Is this just payment for you hosting it? Is there a GitHub?
English
i'm building Kapture - an open-source Loom + Screen Studio alternative
create beautiful screen recordings
share them instantly
free to use locally
pro is only for cloud hosting, storage & sharing
first 100 waitlist users get 6 months of Pro free
English
I think the play is to have multiple tiers of hardware, each for a different use-case.
Instead of focusing on having 1 huge model than can do it all, we can have vision, audio, media gen, agent, recursive problem solving, security etc.. all small and focused.
Cheaper better
English
composer2.5 on pi-codex-goal ran for 12 hours 400M tokens, but the model was never able to update the goal to stop ☠️
thank you @cursor_ai 90% sdk discount
@fitchmultz goated for bringing cursor sdk into pi harness (but idk if your goal extension fully works though)
English
@cursor_ai Just use the sdk as default coding agent for this weekend. On their cookbook clone and run their self hosted cloud agent and now you have dirt cheap composer for 3 days
English
@0xSero @bridgemindai This is my exact setup rn; 2 sparks + deepseek its beautiful. With 1 u can run a less precise version but with 2 its full
English
GPT 5.5 is great to use with my Hermes Agents.
My only complaint is SPEED!
Are there any local options that I can run on my 2 DGX sparks that are fast and smart?
English
@cursor_ai harness turning is nice but the app has a few issues:
1. Worked for time is wrong - this task ran for 2 hours
2. They should distinguish user messages with llm text blocks more clearly
English
@BennettBuhner /goal just keep running - we’re going for a world record
English
they really are true evil
Palantir@PalantirTech
Palantir embraces the neurodivergent. Join us.
English
Did someone really steal my payouts post?
𝚃𝙾𝙱𝙸@tobiflamme
Nobody talks about the hour after 9pm That's when your future gets built or wasted Choose carefully.
English
@cryptopunk7213 The insane advantage they have to fine tune based models based on their own user’s work
English
what an unbelievable come back.
imagine being called “a worthless ai wrapper” for 2 years then miraculously shipping a frontier model that’s
> 20X cheaper than opus 4.7
> 8X cheaper than gpt 5.5
> pioneered its own unique moat (model harness)
composer 2.5 was built on Kimi k2.5 btw (not the newer 2.6).
imagine what this produces when it’s strapped to a 15T model?
Lee Robinson@leerob
Where could we improve Composer 2.5? We're working on the next model and would love your feedback. Lots of work to do (our CursorBench evals below) in the coming weeks!
English
My parents have blessed me with a surreal life and genetic lottery
- 6’1
- American
- Top earner
- Top savings
- Huge tech stockpile
- 2 children + married
- 27 years old
- Top 5% on X
- 1 person away from anyone
- Largest meetup in Poland
- 4 continents / 30 countries
English
