AI News: The New Model That's As Good As Fable

Dude, I've been watching you for awhile and you really are one of the best that cover AI. I don't need to go down the rabbit hole of YouTube. Thank you.

Every test I've seen with Sakana had worse results than just using one of the models it routes to

I don’t know if you heard my thoughts, but your voice sounds so much calmer, softer and nicer in this video — less penetrating, more grounded ☺️ I’ve been wanting to drop this for a while because I love your content and watch you all the time. Usually, though, your voice and tone can feel quite intense, almost like you’re yelling at me the whole time 😂 Even when I turn the volume down, it still cuts through and gets tiring after a few minutes. But in this video, maybe because you’re outside in nature on that beautiful terrace, you sound much more calm and grounded. It feels less overexcited and much easier to listen to. So if this was intentional, congratulations — it really worked 😍

I've been doing "Claude Tag" inside of Discord for four months now. WHERE IS MY YACHT!?

Hey man, I really appreciate your unbiased reporting on some of this stuff. Having a source that just provides the facts without picking sides or jumping into the speculative finger pointing and drama is a big thumbs up.

These views are absolutely mesmerizing, Matt!😍

Great breakdown, Matt. The shift towards 'orchestrator' models like Sakana’s Fugu is fascinating—it really proves that the future isn't about just one 'super model', but about how we effectively route our intent. As I’m building 'Sarthi' (my knowledge hub), this orchestrator approach is giving me some fresh ideas on how to bridge traditional info with AI agents. Keep the updates coming!

Thanks for the relevant updates this week Matt! Love the start/end frame intro, and hope your family is having a great time in Big Bear! 🐻

The Fugu Test was great! Shows what benchmarks don't tell you and clearly shows how good or not so good it is on a few example use cases. That is real value to me rather then just repeating benchmarks from a release post. This is exactly what I'm looking for in videos about new AI models and tools. Please more of that kind of stuff!

Changing up the scene 0:05 was beautiful. Pls do more.