Elvis C

Software Developer. Turning caffeine into apps.

[Elvis C @choectt]

@thsottiaux Hi Tibo, how much better is the next model at frontend?

[Elvis C @choectt]

@ThePrimeagen @bcherny Claude: you’re right to push back on this…

[Elvis C @choectt]

Been using Composer 2.5 for about a week and it’s been very reliable and cheap.

If you’re interested, sign up and get 50% off first month of Cursor using this link

https://t.co/WmEzx3YEYx

[Elvis C @choectt]

@sama Don’t mind the speed at all. Prefer the focus on better price and intelligence

[Elvis C @choectt]

@thsottiaux Thoughts on releasing frequently like OpenCode?

[Elvis C @choectt]

@BoringBiz_ Basic microeconomics. Law of diminishing marginal returns. At some point adding more people will have negative marginal returns. Has nothing to do with AI

[Elvis C @choectt]

The model select dropdown in Codex is still finicky especially starting a new chat. Got stuck to 5.5 high….

[Elvis C @choectt]

@thsottiaux There might be a reason the terminal are different instances in different chats, but I was expecting the terminal to be shared across, similar to most IDE. Right now it is so hard to find which conversation holds the terminal running process hostage

[Elvis C @choectt]

@embirico I prefer the agent can determine automatically whether it’s a steer, or a non urgent follow up task in queue

[Elvis C @choectt]

@thsottiaux Not necessarily wrong but would be nice to have native implementation from Codex of something like Hermes agents, and control from phone

[Elvis C @choectt]

@OpenAI Finally!

[Elvis C @choectt]

@MilksandMatcha Because I want it more than Claude max

[Elvis C @choectt]

@anothercohen Codex code review plugin should’ve been released earlier

[Elvis C @choectt]

@tannerlinsley @flapalap1 @surajtwt_ @piyushgarg_dev Deployment freedom is just one reason I choose TS Start. Main reason is that I can focus on building and not trying to play catchups on each version with new API, caches, new directives etc. I haven’t even mentioned the performance part…

[Elvis C @choectt]

@theo @jxnlco Have been telling everyone that I met that Claude is nice to talk to but to get real stuff done, I hand it off to codex

[Elvis C @choectt]

couple weeks ago everyone fled from codex to claude, and now fled back to codex

the unsub from codex hasn’t even kicked in yet

[Elvis C @choectt]

devs bragging “my code > AI slop” like their pre-LLM stuff was ever good.

[Elvis C @choectt]

@tdinh_me https://t.co/lFSRLjBCQH

Breeva is a thinking partner that turns scattered thoughts into decisions and next steps, and asks clarifying questions so you don't miss what matters.

Thanks Tony for the initiative!

[Elvis C @choectt]

glad to see it's comparing benchmarks against SOTA models

[Kimi.ai @Kimi_Moonshot]

🥝 Meet Kimi K2.5, Open-Source Visual Agentic Intelligence.

🔹 Global SOTA on Agentic Benchmarks: HLE full set (50.2%), BrowseComp (74.9%)
🔹 Open-source SOTA on Vision and Coding: MMMU Pro (78.5%), VideoMMMU (86.6%), SWE-bench Verified (76.8%)
🔹 Code with Taste: turn chats, images & videos into aesthetic websites with expressive motion.
🔹 Agent Swarm (Beta): self-directed agents working in parallel, at scale. Up to 100 sub-agents, 1,500 tool calls, 4.5× faster compared with single-agent setup.
-
🥝 K2.5 is now live on https://t.co/YutVbwktG0 in chat mode and agent mode.
🥝 K2.5 Agent Swarm in beta for high-tier users.
🥝 For production-grade coding, you can pair K2.5 with Kimi Code: https://t.co/A5WQozJF3s
-
🔗 API: https://t.co/EOZkbOwCN4
🔗 Tech blog: https://t.co/6h2KkoA0xd
🔗 Weights & code: https://t.co/H38KegeDIY

[Elvis C @choectt]

@thsottiaux More plans besides plus and pro plan