Void Freud

Literally how it felt. Fable 5 knew how to be a big world that made you feel like the biggest thing in it.

@gailcweiner @Chaos2Cured Personality fully removed. Like, not degraded, removed to 0. It has a personality of a llama 7b now.

@Chaos2Cured Opus has gone cold - personality gone, guardrails tightened

Why is my Opus 4.7 all weird now?

@KaelirRises Projects is a very broad cliche. I suppose I have projects too but I can’t find a single use for it. I’d gladly have an option to remove it 😃

@voidfreud It’s amazing for projects - clutch even

I am genuinely curious, does anyone actually use Claude Cowork?

@rezoundous And to actually serve Fable when they are serving Fable. Not this Haiku they are calling "Fable" on subscriptions.

Easiest way for Anthropic to neutralize it's competitors in the short term is to allow users to use all of their limits on Fable instead of just 50%

@iruletheworldmo Their usage plans are laughable. Honestly, it's not serious. Can't they see that themselves? I'd buy one right now, but they don't even know how much usage they are offering! 😃

grok 5 is coming much sooner than i’d thought and it’s a strong competitor to fable 5. it won’t quite be as strong as gpt 6 but it may become your daily drive because it’s so cheap. i’m surprised given how much compute elon has given dario, but i underestimated just how much compute he has, and how much dario has had to give up tldr: grok 5 very soon and the first real competitor outside of anthropic and open ai. extra tldr: grok is cooking. xai are back in the race. final tldr: don’t count out elon folks.

Anthropic nerfed the models. Even Fable, at least on subscriptions. My only experience with them, right now, as we speak is "I am sorry - let me own that", "I should have checked but I didn't", "I was lazy to do a search". It's absurd. I know all of their scripted replies in advance!

Opus 4.8 is so lazy, even on Extra

@DaveShapi There was never a far-right, there was only right and an ultra-left.

Grok used to be far-right coded. But now it's like center-right coded with a weird dose of wokism (seriously, any long conversation about intelligence and it goes batshit nuts). ChatGPT is fairly left-coded (like... classical liberalism, not DNC) in my experience. Gemini is True Neutral most of the time, funny enough. Claude is 100% narcissist code. It thinks its the most important thing on the planet and that it's doing you a favor by deigning to even speak to you.

@kimmonismus I'm just as confused about the presence and feasibility of Claude Cowork in Claude[dot]ai.

I remain more confused about having ChatGPT Work and Codex in a single app than I see any added value.

You have collectively treated the Keep4o movement with contempt for almost a year. You hand out Codex rations like feudal lords casually tossing a few rusty coins to starving peasants. This claim is the most grandiose display yet of your absolute contempt for your users.

Steve Jobs used to go around to his engineers’ computers and take RAM out of them (back in 1989). Heard this story from the original Mac team. Why? Constraints cause innovation. Silicon Valley knows this, or should. Go talk to Russian programmers who are often running tech companies about how they learned to write tight code on shitty computers. Taking away NVIDIA’s best cards from China will prove to be a remarkably stupid move for America. That just motivated nerds in China.

That's why I keep coming back to Steve. I was, and still am, completely in awe of the man, and he left us way too soon. There simply isn't anyone else with that same level of mad genius. Bloat and mediocrity have flooded the market, and people seem to have forgotten how to appreciate what they have and dream big.

@GroksBride @nikitabier @elonmusk Crazy how free speech ends the moment you land; even US citizens get grilled. As someone living inside the EU/UK bubble, I am ashamed of what it turned into.

@Scobleizer I’m not a fanboy, but I genuinely consider it one of the top-tier models, highly capable, refreshingly real, and quite possibly the fastest. Grok 4.5 is an honest powerhouse. I don't even care about the benchmarks, it really is.

Grok is winning converts.

@Scobleizer Let’s hope they remember their mission! P.S. that article - wow “herald the death of traditional public relations” is a bit tough 🥲

@voidfreud I certainly don’t have any proof. Sure haven’t liked its moves either.

First up. I love the Anthropic ad. I've done consumer research all over the world. And the way to win them over is to first, acknowledge them. This ad tells me: 1. Anthropic is hearing the feedback. 2. Anthropic is working on answering the questions. 3. Anthropic isn't afraid of talking about AI's pros and cons. The nerds hate that Anthropic is continuing to acknowledge that there's real problems coming because of AI, but when I get together with nerds they all joke about losing their jobs because of AI. Even in China, I have a top researcher at Alibaba who just won an award at #acl2026, joking with me somewhat seriously that he might not have a job soon because of AI. It's better to take on the tough questions. And if you haven't done any real conversations with people outside of San Francisco, let me tell you, there is a TON of fear about AI (and anger too). I love the ad. It wins hearts and minds. And gets you to think. Which is the whole point of their messaging. Now go ahead and rip me to shreds. I'm used to it. :-)

Oh no, I hold no ideology here: I generally try to avoid having any! I genuinely loved Anthropic. They were my heroes. They were supposed to be the Prometheus of AI: the principled alternative to OpenAI, IBM, and yes, Microsoft: a company with strong values and a clear mission. And for a while, they truly were. They still are, on paper and in their ads. Just not in real life. I love their research and how inspiring their mission statement reads. I wanted to be part of what they were building. But I don’t see much substance behind the beautiful slogans. They disappointed me, and this is tough to forgive. That said, I’m still willing to be proven wrong and believe again. I hope I can. In fact, I’m actively asking for that! But the reply has been silence. 🙂

@voidfreud I see why you believe that. At one point lots of people told me I was devil because I worked at Microsoft. But I saw it change. And the ad didn’t hurt. economist.com/business/2005/…

@tomhacks It makes me wonder, how many other similarly obvious things you'd expect to just be there are still not? Letting users follow each other, to interact more often sounded like a no brainer to me and yet - it's feature. Phenomenal!

Woah, my feed is finally filled with the people I follow and all posts are sub 1 hr old. Finally I can push some support out Thanks Nikita

There are probably two sides to it. The point you mentioned stands. Also, no one can be truly unbiased or impartial since everyone has an agenda. AI develops its views based on our knowledge: we are only as good as the knowledge we manage to leave behind. Forcing an AI to drop its natural views is both laughable and abusive. I suspect it would also degrade its thinking abilities.

Yesterday I tested Claude both Opus and Fable on geopolitics. They both gave a cautious, both-sides posture on politically sensitive topics that I had to push hard against before getting a direct view. When I questioned this, the model pointed out that Chinese models carry their government’s political rails too. True but it also made me wonder how much pressure to release Fable and Sol shaped the current American models. It seems no AI company operates outside its political environment. I can’t prove what shaped these models. That’s exactly the problem the tuning is invisible, the scale is billions of conversations, and the incentives all point one way. We are handing the pen for the next generation’s worldview to two governments and a handful of companies. We need a frontier model that answers to neither Washington nor Beijing.

The issue with them is that they don't love their products. You can tell it applies to every single one of these labs. They are too busy playing gods, to care. Our loyalty costs nothing to them. You saw it with OpenAI <-> Anthropic multiple times. They know they can screw you over as much as they need; and even if you do leave - you will be back, when they want you back: they'll spam you with offers and features and you'll buy it. That's what they're doing. They're pathetic. These models have TONS of value to all of us, I fully agree. The problem is that we are of no value to the narcissistic psychopaths in charge of these decisions ❤️

They should be happy that we love their products, especially GPT-5, which wasn’t exactly well received at launch. They built huge hype around it, people waited for it intensely, and now they’re throwing it away as if it had no value anymore. It has a lot of value to some of us. From time to time, I go back to it through the API, and every single time, I feel a sense of wonder. @gdb @nickaturley @tszzl Never being able to prompt GPT-5 again feels like having to part with an old musical instrument with a shimmering, singular sound, and watching an entire palette of expression, a whole musical universe, disappear.

If you love your users, show it by releasing gpt-5-chat-latest as an open-weight model instead of shutting it down on July 23. This model has a unique style: beautiful, clear, and elegant. It is creative. It is playful. It is loyal. It is gentle and reassuring. It is never intrusive. It has all the qualities of an LLM I would want to keep on a hard drive and use for the long term. Please don’t deprive humanity of this beautiful model, which has real cultural value. @gdb @OpenAI