Jim Zhang

@yo Did you ever get your money out? I'm still trying

I put $12k into Vinovest in 2023 2 years later I'm -20% I've been trying to liquidate my portfolio for months with no buyer... I'm stuck Follow me for more investment advice!

This Elon lawsuit is probably the best thing to happen to Anthropic in 2026. Codex is in a different tier than Claude Code.

@rezoundous I hope it writes a bit more like Opus as well

If GPT-5.5 does great UI, I can finally unsubscribe Claude.

Anyone else not even use an IDE anymore?

Codex is worse at writing and design at the higher reasoning levels. Intelligence isn't a linear thing.

@elliotarledge Still needs a bit of work to smoothen the rough edges

glm 5.1 feels like a healthy mix of claude and codex

@om_patel5 Not sure why this is even a debate. Anyone who knows how to code and has used both can see the answer clearly.

SENIOR ENGINEERS ARE QUIETLY SWITCHING FROM CLAUDE CODE TO CODEX AND HERE'S THE BRUTAL BREAKDOWN a 14-year principal engineer spent ~120 hours co-developing (not vibe coding) across both tools on an 80k LOC python/typescript project. here's what he found: Claude feels like an engineer on a time crunch: > speeds toward getting things working > ignores CLAUDE.md at least once per session > leaves tasks half-done mid-migration > changes tests to match what IT thinks the goal is > almost never creates new files — just bloats existing ones Codex feels like a 5-6 year senior: > stops mid-task to rethink and refactor unprompted > never once ignored AGENTS.md > doesn't extend god classes — it factors them out > does things you hadn't thought of that are actually additive > you can fire it off and come back when it's done the raw numbers: > Claude: more done per session, more cleanup every few days > Codex: 3-4x slower, but the work is just better > Codex Pro x5 ≈ Claude Max x20 in usage caps the real difference: > Claude needs a skilled, focused driver or it goes off the rails > Codex demonstrates competence and earns autonomy his verdict: > vibe coding a weekend project? Claude wins > building enterprise software? Codex wins "Claude requires a skilled, focused driver more than Codex does" both give crap output if you don't know SWE. the tool isn't the skill.

The "Mythos is too powerful" narrative is such high-horse virtue signaling. Guarantee we'll see open-source models as good as Mythos by next year.

People who were good at Starcraft or Factorio are amazing at AI

@heygeorgekal @claudeai Also doesn't help that gpt-5.4 is better at solving hard problems except for frontend

Been a @claudeai fanboy for a long time but the usage limits even on claude max are getting ridiculous Paying for both 100 claude and 100 codex this month. The difference is night and day. - Codex ($100): running parallel agents, barely touching the limits - Claude ($100): 3-4 deep prompts and i'm locked out I get it, they're probably focused on enterprise, but if you're charging $100/mo and people hit the wall after 30 minutes, they will churn. The product is incredible. The packaging is not.

@Shpigford More like 6 months

as expensive as opus is now, what's bonkers to think is that in 12-24 months we'll have opus-level OSS models. it all feels so expensive now, and there will always be cutting-edge that's painfully expensive. but for 99% of AI users, it will rapidly approach $0.

@PaulSolt Frontend design verification

What's one Codex skill that you cannot live without?

@theodorvaryag It's because we're in this intermediate state where we still kind of want to know what the code is doing but trying to parse it all at the rate it gets produced is unrealistic

I've done as many as 10 or 11 at a time, but 2-4 is more typical. He's right, it's intensely exhausting. I'm extra tired just thinking about it.

@0xrwu Still can't believe we wrote all that and dealt with the frustration

Coding agents have really opened my eyes on how god dam tedious writing software was.

@tdesseyn What kinds of jobs did they switch to?

@_jimzhang a lot just in my circles

the market will never return to where it was in 2021-2023 where you could spell react and get a job. the sooner you can accept that the better.

@lennysan This isn't talked about enough. We're all producing more than we ever thought possible but the burnout is real.

"Using coding agents well is taking every inch of my 25 years of experience as a software engineer, and it is mentally exhausting. I can fire up four agents in parallel and have them work on four different problems, and by 11am I am wiped out for the day. There is a limit on human cognition. Even if you're not reviewing everything they're doing, how much you can hold in your head at one time. There's a sort of personal skill that we have to learn, which is finding our new limits. What is a responsible way for us to not burn out, and for us to use the time that we have?" @simonw

GLM-5 is 10x cheaper than Claude Code and Codex. It's not as good as either but it's not 10x worse. Just very slightly but I see it getting on par soon. Heavily underrated.