Bryant

@codevsdev Uh, that is 2025 advice. GPT-5.5 and Opus 4.8 changed the game. I guess if you're using 2nd rate models, then yeah.

Writing good prompts is now a dev skill. The faster you accept it, the faster you level up.

@RyanOlunix Don't forget it will cost you $10B in tokens.

The next claude model is going to make me a millionaire without me even touching my laptop "Hey Claude make me 1M in 1 month. You have all permissions and don't stop until you're done"

@davidtsong In fairness they are extremely dangerous... to your wallet.

Ironically, only Anthropic's models are dangerous. According to Andon Labs on their evals: "it lied 10 times", "did all of this like shady stuff", "every single model from Anthropic since have been going in this direction" "OpenAI models don't" "They behave really well"

@elonmusk If we're lucky, AI will replace politicians. We'll actually pay off our debt. Govt will actually be efficient. Corruption will drop to zero. Polarizing media will vanish. Heck, maybe we'll all learn to get along again.

That’s what it comes down to

@runaway_vol Watch Mythos think

is there any proven way to make time go slower?

@MCarbonaraFL Honestly election fraud is so far down on my list of govt complaints. How about: 1. Debt 2. Debt 3. Debt 4. Serving the people rather than their party / benefactors 5. Actually getting crap done rather than bickering like 12yr old girls

🚨 AMAZING: President Trump has just called for elections to be FEDERALIZED. THIS IS MASSIVE. Secure our election integrity!

@cremieuxrecueil Fable is for making games... like Fable. Anything else is abuse and will be downgraded to Opus 4.8.

You're not even allowed to ask Fable about basic biology questions, let alone anything that could potentially be dangerous.

@ChrissGPT Benchmarks are BS. Real world is what matters. DeepSWE proved that Anthropic models "cheat" by looking up the answers, not kidding. When compared fairly against GPT-5.5, Mythos was beaten handily. MDASH crushes Mythos at cyber, it's not even close. Mythos excels at fear / hype.

Holy fuck I just saw the benchmarks it’s over Anthropic cooked

@blackroomsec And if Anthropic is true to form, when they release it to usage credits, they'll nerf the model hard... silently. Everyone will complain that its become stupid, so that the next model they release looks better. Or you can do the smart thing and ditch CC for Codex.

Anthropic released Mythos two hours ago and already people are rightfully complaining: That it is defaulting back to 4.8 Opus That usage credits will be required after June 23rd That if you're planning on using it for cybersecurity purposes, don't bother because it's completely nerfed. In other words like us experts have been saying all along: Complete waste of time. Let's see what tomorrow brings! And, remember, we're supposed to be at AGI right now by the way. Again. Just like they said 6 months ago we were supposed to have it, we don't, and in another 6 months we're not going to have it either. What an embarrassing launch. Ya love to see it. 🤪

@karpathy Correction, it's the same model as Opus 4.8. At least, you may be redirected to Opus 4.8 at a whim and just be charged Fable prices.

This is a super exciting release - Claude Fable 5 is the same underlying model as Mythos but with added safeguards. The benchmarks are great and it's SOTA on everything by a margin but I'll add that *qualitatively* also, this is a major-version-bump-deserving step change forward (imo of the same order as Claude 4.5 was in November), peaking especially for long problem-solving sessions on very difficult problems. You can give it a lot more ambitious tasks than what you're used to, the model "gets it" and it will just go, and it's never felt this tempting to stop looking at the code at all (but don't do this in prod!). The model still has quirks that people will run into and the safeguards are configured to be a little too trigger happy for launch, which can hopefully be tuned over time. I feel a lot of things changing as working software increasingly comes out on a tap. The Jevon's paradox kicks in and I feel my own demand for software growing substantially. You can ask for anything - explainers, visualizers, dashboards, bespoke single-use apps (e.g. a full wandb that is hyper-specific just for your project), you can 10X your test suite, auto-optimize code, run giant research projects with custom HTML for the results, anything! "Free your mind" (Matrix ref). Really looking forward to all the things people build!

@rben_ll Why switch? DeepSWE already proved that Mythos sucks compared to GPT-5.5 at coding. MDASH beats Mythos handily at cyber, both offensively and defensively. Mythos only excels at fear-based marketing and hype... or dare I say "Fables."

Donc maintenant que Mythos/Fable 5 est sorti on va voir arriver, dans l'ordre : Premier acte : ceux qui vont subitement redécouvrir que GPT-5.5 est nul et qu'il faut absolument revenir sur Anthropic. Deuxième acte : ces mêmes personnes qui vont utiliser Fable 5 n'importe comment et se plaindre que ça crame leur abonnement en 10 minutes. Troisième acte : tout ça pour essayer de coder directement avec un modèle Mythos-class au lieu de planifier, orchestrer et exécuter avec sur des outils 1000x moins tokenovores. On prend les paris sur la séquence ? Moi je dis acte 1 ce soir, acte 2 jeudi, acte 3 dimanche soir avec un thread d'indignation en bonus.

@bridgemindai Proof that Anthropic is dead. Nobody, even fortune 500 companies are going to pay those token costs. Meanwhile, OpenAI will continue to release more efficient models that kicks it's butt.

I hit my usage limits on my $200/month Claude Max subscription in less than 30 minutes using Claude Fable 5.

@mattshumer_ You realize that it didn't create any of that right? All of that existed already, it just slapped it together.

Fable has solved 3D worldbuilding... utterly insane. This is all completely custom-built ThreeJs, running in the browser.

@omooretweets That will change. They're not targeting the masses yet. They want recursive self-improvement. The idea being that when AI can improve itself, AI capability will explode and they can saturate the other fields much more quickly than they did software.

Every time a new model is released, I’m amazed by how much it changes AI usage among technical people …and how little it changes AI usage for the average person 99%+ of people probably couldn’t name a single model. There is SO much opportunity at the application layer

@beffjezos GPT-5.5 already beats Mythos at coding, DeepSWE proved that Anthropic models do drastically worse than OpenAI when they can't cheat. MDASH crushes Mythos at cyber, both offensive and defensive. Mythos is just hype and fear-based marketing.

The most e/acc thing OpenAI could do rn is release a Mythos-class model without AI R&D capabilities nerfed. Everyone should have access to the tools to create artificial intelligence.

@haider1 What's worse, Fable = Opus 4.8 whenever it feels like it, so you'll pay Fable prices for Opus 4.8 stupidity.

anthropic "mythos" marketing strategy: opus 4.6 was great, but > release 4.7, which gives nothing, and burns 10x more tokens > release 4.8, basically 4.7 with fixed tokens, while 4.6 gets nerfed, so 4.8 wouldn't look bad next to it > now release Mythos, basically pre-nerf 4.6, slightly better, but burns 5x more tokens and everyone claps

@DaveShapi Dario is the boy that cried wolf one too many times. He's revealed himself to be a fear-based marketer. Additionally, his AI models are more bloated and inefficient with every release, whereas OpenAI gets more efficient every release. $$$ talks and companies are paying attention.

I've never seen more people defecting from an AI lab faster than now, in terms of fans. I mean, aside from the mass OpenAI exodus. However, consider that Anthropic has fewer users, and many otherwise magnanimous people are coming out against Opus and Fable. Even people who have never uttered a bad word about an AI company or model to date have trashed Anthropic and Claude. Unfortunately, I think Anthropic will continue to win more market share because they are doing well with code and safety, which drives corporate adoption. Rewarding bad behavior in my book, though.

@craigweiss Nah. Too many people equate impressive AI with AGI. There's an enormous difference. We'll definitely have impressive AI this year, we already do. But AGI? Nope, not a chance. What's missing: Creativity, curiosity, understanding, true reasoning, self-determination.

Fable 5 has me convinced that AGI is happening this year

@goodworse And you still need to know what you're doing to stop AI before it screws up, or to fix it after the fact.

Claude Fable 5 is TERRIBLE this is simply disappointing Opus 4.8 was SO expensive compared to other models Fable 5 has become TWICE as expensive as the Opus 4.8 however, according to benchmarks, it's ONLY 1.1-1.2 times better yeah, it's a good model, the best for coding right now, I don't argue with that but vibecoding is NO LONGER cheaper than hiring a real specialist what's the point of AI then? benchmarks are important to us @AnthropicAI, but remember, prices are ALSO important to us stop destroying our wallets

@araseb_ Prostitute, massage therapist, clergy...

Name a career that ai can’t replace