From: Vladik on 𝕏: https://x.com/Kostoglodov/status/2071144065857679631 Shaw (spirit/acc) on 𝕏: https://x.com/shawmakesmagic/status/2070918006033817867

Literally everyone in tech keeps saying AGI (Artificial General Intelligence) is coming in like 2 to 3 years. But honestly, looking at how things are going right now, I’m starting to doubt it. Is true human-level AI actually inevitable, or are we just hyping it up too much? Here is why I feel like it might happen: - AI is building AI: Models are already writing a ton of their own code now. If AI just keeps upgrading itself, it's bound to explode at some point. - Reasoning models: AI isn't just spitting out the next word anymore. It actually pauses and "thinks" through math and coding problems now, which is kind of insane. - Infinite money: Big tech companies are spending hundreds of billions of dollars on data centers and chips. They aren't going to stop anytime soon. But here is why I think we might hit a wall: - It still feels fake: An AI can solve a crazy math problem, but then fail at a super simple logic puzzle just because it wasn't in its training data. It feels like super fast memorization, not actual smarts. - Power limits: These things use a terrifying amount of electricity and water. We might literally run out of power grids before the AI gets smart enough. - Moving goalposts: Every time AI does something cool, we just say "okay, but it's still not real AGI." We don't even know what the finish line looks like anymore. Are we actually going to see sci-fi level AI sometimes in 2027-2029, or is this whole boom about to plateau hard? Drop your thoughts below.

The consensus is clear, and it's not even a debate: Fable 5 was a monumental leap in capability, and the community is now in a state of collective grief now that it's gone. Long story short, Fable is to Opus 4.8 what a senior architect is to a junior dev. It doesn't just write code; it plans, researches, and executes entire complex projects with minimal handholding, often correcting the user's own flawed premises. The thread is full of mind-blowing examples of what people accomplished in the short time it was available: - Insane Coding Feats: Users one-shotted entire games (a 3D animal hybrid game, a near-perfect Dungeon Master 2 fangame recreation), built complex apps from scratch in hours (Outlook add-ins, mobile stat-tracking apps), and had Fable autonomously refactor and improve entire SaaS codebases. - Next-Level Reasoning: It wasn't just for devs. Users had it analyze their golf swing from a year's worth of videos better than a pro, crunch billions of lines of horse racing data, and compare complex legal documents more effectively than an expensive law firm. - Autonomous Agency: A key theme is Fable's ability to use tools, manage sub-agents, and maintain context over long, complex tasks without getting confused. It could plan and execute a multi-week project backlog in a single session. The overwhelming sentiment is that going back to Opus feels like a painful downgrade, with many calling it "a caveman" in comparison. However, one user did note that for non-coding tasks, it felt more like a minor improvement than a total revolution. So yeah, the hype was real. Now everyone's just sitting here wondering when we'll get our super-intelligent partner back.

Back to posting here. Sorry it's been a while. Most of us reading this post will never get to use Mythos, but luckily Anthropic decided to release a version with a ton of safety guards (Fable) The benchmarks look strong, especially the jumps in vision and code. Curious how Gemini 3.2 and the next GPT models stack up. Working in the Risk & Compliance space, my daily prompts deal with fraud and money laundering. With the new safety filters, and since I'm not in Anthropic's special program, those exact queries may get rerouted to Opus 4.8 or blocked outright. The tool built to catch bad actors might end up flagging the people fighting them. I broke down the full release, press notes, and system card in a new 20-minute video https://www.youtube.com/watch?v=TtCWqjy3T5M&t=3s

I was already aware that I would be reverting back to Opus 4.8 when Fable 5 would move to Pay per usage. By chance, I had it writing skills for Opus 4.8 about 10 or so hours before they were made to pull the plug and I've had these skills enabled for the past weeks working through my personal projects on Opus 4.8. Perhaps it's cope, but I truly believe Fable knew "how Opus 4.8 thought" and could help point it's thinking in the right direction with these skills. Please note that I can not 1. prove these skills actually helped 2. share these skills since they were also mixing things in relating specifically to my project I nevertheless think it's worth people do this now! And I'm also interested to here if you did something similar yourself / you think this is nonsense (and why?)