The most powerful consumer AI model ever has been released today

[LordFall]

This is the training wheels version of Mythos that got Anthropic to urgently gather up 40 top institutions to warn them about it's capabilities and lead to Project Glasswing. 

The models at this point have the power to break our current understanding of cybersecurity and they also warn about biology for fear of bioweapons being created. On the flip side that means major medical breakthroughs are coming as well so if you're suffering from a previously incurable illness like autoimmune disorders or cancer and what not then it's looking better than ever.

Buckle up boys and girls, the next 5 years are gonna be a ride! It's only accelerating. 

Edited Tuesday at 06:34 PM by LordFall

[Natasha Tori Maru]

What is actually new? Like pragmatic results.

Not hearsay 

Did it crack the IRS database?

[LordFall]

Fable 5's greatest leap forward is agentic workflows. They can work for hours up to days apparently. Agents usually lose coherence quite quickly so they aren't able yet to be a reliable employee. 

The original version of it Mythos is apparently great at finding zero day exploits and basically can be set to hack random software en masse which is why Project Glasswing was established. 

[Apparition of Jack]

With all these new AI models I feel like I’d be cheating on my bae chatgpt 😭 😭 

[LordFall]

1 hour ago, Apparition of Jack said:

With all these new AI models I feel like I’d be cheating on my bae chatgpt 😭 😭 

Yeah it's fair to find your comfort model. I like Gemini the best. I just started playing around with Claude. It's way more argumentative than Gemini so that's good for pushing intellectual boundaries. It can also be installed directly on your computer to play with your files directly which freaks me out a bit so I won't do that for now but I'm sure it's useful. 

[NewKidOnTheBlock]

It's kinda exciting getting these kinds of news, not gonna lie, I'm always curious how well they're doing compared to the previous versions from various companies on variety of benchmarks, and this one is doing very well indeed🙂 I hope they continue to improve, if for nothing else, just to witness all the luddite bitterness😀

[LordFall]

Agreed! I'm most excited that they're very fearful of the biological capabilities of this model and have gone overboard to censor it. To me that means that once they scale this out in vetted universities and researchers the level of medical advances that will come out in the next few years will completely change our paradigm.

I'm a big believer that we will reach biological immortality soon and this is a proof of it.