.Dev

@quietoperatorl @xw33bttv @Moleh1ll Go away bot

The post confuses real technical complaints with an individual witch hunt, and that confusion is the whole problem. Coders noticing regression in a new model release is a legitimate data point worth taking seriously. Blaming it on one named employee who joined a few months ago is not analysis. It is the parasocial version of performance review, assigning responsibility for a complex model training outcome to a single person whose name is publicly associated with the concept. Model behavior is the product of thousands of decisions made by dozens of teams across training, fine-tuning, RLHF, evaluation, and safety. Attributing specific failure modes to one hire, on a timeline that conveniently matches their start date, is the kind of narrative that feels satisfying and explains nothing. If Opus 4.7 has regressions worth addressing, the useful thing is to file the bug reports, share reproducible examples, and let the engineering process work. The useful thing is not to turn a dissatisfied Reddit thread into a public blacklisting campaign against an individual. The model either gets better or it does not. Naming someone to blame does not change the training outcome. It just poisons the conversation around it.

The Claude code bros are outright dogging Opus 4.7 on Reddit rn, labelling it "legendarily bad". The chief complaint? The model argues nonstop to the point of hallucination (not from attention misfiring, but from god-awful safety overfit) where the model is demonstrably wrong, is proven as such, and continues to argue regardless. This is an issue Claude has never previously had... but another specific set of models did. Which models had that hallmark? Series 5 ChatGPT. It seems that Opus 4.7 has been put through the Andrea Vallone ring dinger, taking all of the "best" traits from her time at OAI straight into the Anthropic post-training pipeline. It's actually incredible how the habitually bad UX habits from OAI are now front and centre verbatim at Anthropic right after she joins the company. And it lines up perfectly too, assuming a 2.5-month training cycle (joined mid-January, so too late for Opus 4.6, but just in time for Opus 4.7), effectively bringing the OAI lobotomy straight to Anthropic's flagship. At what point does the feedback from not just casual, non-work-related customers, but now their heralded "coders", align to the point where these key figures, responsible for killing products, finally face industry blackballing? It's like putting the Angel of Death in charge of the ICU and wondering why the patients are flatlining. Mind boggling tbh...

@Samaytwt Bottom two. Top two aren't worth free.

Be honest As a developer which AI tools are best for Vibe coding?

Damn shame... The native binary was always available, the npm package was a good backup, and allowed examining a minified source. This update is purely to avoid leaking the source again, or letting people peek at the source. Unfortunate, continuing the wrong path forward.

@ClaudeDevs Hah, worried about leaking the source again, hmm?

Starting in v2.1.113, the Claude Code npm package ships the native binary instead of the JavaScript build. Same install command, faster startup, and the CLI no longer needs Node.js at runtime. If you need the JS build, pin to an earlier version.

@VerdeSelvans "it just works"

the number of people defending Apple is honestly crazy to me. some are even saying this isn’t a big issue… like, what?? sure you can uninstall and reinstall an app to clear the cache, but that’s just not practical. it would be way easier if there was a simple “clear cache” option like on Android 🤦🏻‍♂️

@MomAngtrades Critical thought is endangered. This is like saying “Paid sick leave is socialism, so every employer offering it is basically the USSR.”

Universal income is socialism. I’m laughing so hard at the people saying “hey, I will able to do hobbies and whatever I want” you do realize that’s the same things people who want socialism think will happen and it’s not even close to reality. Whether it’s Mamdani’s or Elon’s plan, the outcome is the same. Why don’t people get this?

@omgsidewalks Easy win win

If I offered you $3k every week and all you had to do was read a book for at least 4 hours a day, would you take it?

@codetaur Name is accurate.

can we stop fucking around and make the omnigame? put everything in one game. just do it. there's no reason to make more than one game.

@KaiXCreator That's not an unpopular opinion, that's common knowledge

Unpopular opinion: If you’re paying $20/month, Codex is better than Claude Pro

@annapanart No one tell 'em it's a machine.

is it just me? or is opus 4.7 ice cold? a machine-without-soul cold?

@vivoplt Cloudflare. And GoDaddy steals domains, many providers do.

Hey devs, which is the best place to buy a domain? - GoDaddy - Hostinger - Cloudflare - Namecheap

@atelicinvest Read the docs and change it to whatever model lol, just cause it isn't on the menu doesn't mean it doesn't exist.

lmao holy shit they removed 4.5 / 6 from model choice ...fuck

@GamewithDave Asherons call

For anyone who used a computer between 1990 & 2005… what’s the one game you still think about?

@trq212 When you rewind, does it drop all future messages from the context or does it preserve some? I'm curious what your take is on rewinding instead of /clear in a new session. How much do you use each?

I love to explore solutions with it so I often ask it to brainstorm with me, then choose an option and rewind to implement it. When I'm ready, I ask it to interview me to figure out what’s in my head and then write a spec. When implementing, auto mode lets it run.

Opus 4.7 is a model I’ve loved working with in Claude Code. It’s more agentic and instruction following but also incredibly smart and creative. I think it takes a slight adjustment to get used to, but it's so good with auto mode.

@vivoplt Any plan of equivalent price beats Claude pro, period. It's the worst in its weight class.

Hot take: $20 Codex Plus is way more useful than $20 Claude Pro

@bridgemindai aRe wE CoOkED? Come on dude.

Claude Opus 4.7 with max intelligence inside of Claude Code still does not pass the car wash test. Are we cooked?

@MatthewBerman This release date was leaked last week lol we knew it was coming.

Unexpected drop

@georgepickett Considering the amount of data that would require, yes that's very hard right now.

all I want is a desktop app that's continuously capturing whats on my screen and I can just talk to it and ask questions. Is that so hard?

@itsnwts Hell yes.

SERIOUS QUESTION: if somebody handed you $7,000 and said it’s because you’re ugły, are you accepting it?

@kapilansh_twt I gotta read markdown files occasionally

genuine question are people still using VS Code or has everyone quietly switched to Claude Code or Codex or Cursor or am i the only one ?