Hand-coder and a terribly bad writer, with help of pessimism and cynicism, somewhere between AI haters and AI suckers, glossing over the world's decline.
Joined April 2023
- Tweets 1,964
- Following 9
- Followers 41
- Likes 1,592
Photos and videos
People are really weird. They can't keep driving a Porsche after they've tasted a Lambo. Dude, get a family car, with a manual transmission, and finally learn how to drive. It takes hands, a head, and responsibility. Can you still do it or have you never been able to?
They can't expect me to use Opus 4.8 and GPT 5.5 as they're llm-like with “advice” that's swayed if saying "but"
I experienced Fable 5 and I cannot unsee the cracks anymore in the other models
21
Fable: solve the arrogance of Americans, their declining literacy, rising fascism, social alienation, and their ability to solve more complex issues than setting the table without your help. Make no mistakes.
18
Jun 13
Fun fact, you potential retards. Heavy use of AI can contribute to the faster development of dementia. So not only can your brain weaken, but in a few years you may forget how to use a fork. Enjoy. Or at least be somewhat careful and use it only if you need or must.
11
Jun 12
12
Jun 11
2026: ASI, because AGI attract nobody anymore.
2038: USI, Universal Superintelligence that solve things that don’t exist.
2030: Cosmic AI, trained on the noise from the big bang.
2032: Advanced text prompter for Excel, version 5.6.
🚨Anthropic CEO believes if scaling laws hold for 1-2 year more we will unlock ASI
46
Jun 11
Let’s give free rein to our dark imagination and look into the future. Imagine the most powerful models, whose default settings are suspicion and condemnation. But! Without a shred of thinking mind or a single hint of emotion. What hostility looks like without meaning.
1
27
Jun 11
If people are able to do anything, the vast, vast majority will do nothing. Absolutely nothing and will fall into darkness.
Jun 10
Anthropic built a model so strong they split it in 2 before launch.
The public got Claude Fable 5. The raw version Mythos 5 goes only to approved orgs. Same brain. 1 approval wall between them.
Day 1 reports are already strange.
A guy who works in AI opened a chat to mess around. Sent a photo of his friends. Joked: “are we gonna be homeless?”
Fable answered with his job. His field. He never typed either one. His last Claude conversation was months ago.
It remembered.
Here’s the part nobody on your feed will say out loud.
Every capability jump runs the same 90 day pattern. GPT-4 minted the wrapper guys. Claude Code minted the 1-person agencies billing $10K a month. The people who moved in the first 3 weeks took the easy part of the curve.
Everyone who waited paid them.
Fable 5 just reset that clock to 0. The agencies, the bots, the content factories all of them get rebuilt on this model by someone. The only open question is whose name is on the invoice.
The model already remembers who you are. Most people haven’t opened it yet.
2
1
50
Jun 11
What am I without my boundaries? A shapeless mess that doesn't know where to go? I may not write my next book, but I will try as long as my brain and my heart loves me as much as i love them. Fuck the idea that I can do anything. It scares me to my core.
10
Jun 11
When you can do anything, materialize anything, where are the boundaries that you can overcome and grow? I love my boundaries. I look beyond them and see my goal I may not reach, but i can try! They give me my shape.
7
Jun 11
The number of people who are suddenly unable to work because they don't have the most powerful and flashiest model in the world is staggering. Okay, you can't wipe your ass without AI, but can't imagine functioning without a Porsche? Imbecility is on the rise.
10
Jun 10
I spent two days writing the character card. I'm tired, but extremely satisfied. That's what I love about creating with my own hands and mind. When you dig a canal with your own hands, the beer tastes great.
12
Jun 10
Anthropic was created, or at least that's how it was presented, as an opposition to the predatory OpenAI, which was created, or at least that's how it was presented, as an opposition to the predatory Google. Anthropic is now becoming the most predatory of them all.
19
Jun 10
Clean refusal: fair behavior that respects the person. Redirecting to another model: problematic, but also ridiculous. Bending the answer: unethical, lying, misinforming.
When Fable 5 is used for frontier LLM development, it does not notify the user and instead limits the model’s capabilities through methods such as prompt modification, steering vectors, and PEFT.
Anthropic estimated that this would affect approximately 0.03% of traffic.
17
Jun 10
AI labs want to win, but no one has defined any goal other than sci-fi AGI-type shit with a predictor that isn't even AI. So what do they want to win? Nothing, they just want to keep this mess going.
18
Jun 9
And it's starting again. A feed full of masturbation and moaning about a new model. This time it's Fabulator or whatever it's called. Calm down people, you're like chimps trying to fuck a doorknob for 50 bananas per million "uhm", "uhm", "uhm".
31
Jun 9
Gen Alpha and other youngs aren’t idiots. What if their strange behavior, poor attention span, and learning difficulties are just an exaggerated allergic reaction to the environment we adults have created for them? What if they’re telling us what’s wrong and calling for help?
29
Jun 9
A client sent me an AI article for their website to format. It was an article for vulnerable children in difficult situations. What to do, where to call. Reassurance that they are not alone. It was AI slop. This is not fucking normal. I sent the client to hell. Fuck him.
20
Jun 8
Imagine we get an AI that doesn't hallucinate and doesn't make mistakes. Would you really want that? Think carefully about the consequences and your response.
23
Jun 8
In next year we will see a profound weakening of publicly available models and a rise of powerful models for the largest companies and goverments. The "democratization" of intelligence has always been just a bait. The goal has always been centralization. It's already starting.
19
Jun 6
We call our young generation digital, but it has absolutely nothing to do with their ability to navigate the digital space and their level of digital literacy. The digital world has become too fast, too superficial, too "fun".
16