@slash196
DeepMind solving protein folding and making matrix multiplication faster is one short of the insane breakthrough hattrick.
@liberty-matrix
"The greatest shortcoming of the human race is our inability to understand the exponential function." ~Al Barlett
@GoodHorse413
This is some of the most unbelievable science news I've ever been alive for. It's absolutely astonishing. I can't help but think that I'm only 23! I can't even begin to imagine what the world will look like when I'm in my 40's.
@alansmithee419
"Matched our best human's efforts in 75% of cases..." Oh so it only failed in 25% that's pretty good. "...and exceeded them in 20% of cases." !?
@eduardofigueiredo3113
Correction : The winograd method could already multiply 4x4 matrices with 48 multiplications since 1990 the breakthrough here is that this one does it over non commutative rings
@lucaseev
this could be the inflection point that everyone has been expecting
@andyp123456
Demis Haasabis is the only AI CEO I actually like. Seems like a nice guy, plus he helped create the game Theme Park at Bullfrog in the 90s, so he's gotta be one of the good ones. Definitely doesn't seem slimy like most of the others seeking their 400 billion dollar IPOs at the peak of the hype, and clearly is in it to solve difficult problems to make the world a better place. Bravo Deepmind!
@jumpersfilmedinvr
Okay bro. This is a significant gift they've allowed to go public. This is the level of evolutionary maturity I'm proud to be a part of lol.
@CuteCabbageMonster
I'm hyped but also a bit scared
@Holammer
AI improving itself, that's the first steps in what futurists call the technological singularity.
@lawrencefrost9063
So it begins..
@igoromelchenko3482
What a time to be evolving 😅
@Highkingcuddles
It's insane how fast everything evolves. I'm shocked every few weeks or months now, not years.
@jamieyoung3770
Well, I do hope that these very intelligent systems we are developing will find a use for us someday.
@ThyCrowing
I, for one, welcome our new AI overlords.
@Rebel101
It seems we are at the point where AI algorithms and designs may no longer be understood by a wetware. What could possibly go wrong?
@machinegod3000
"This self-optimization is currently constrained by human-defined tasks. But imagine a system allowed to improve itself without such limits — autonomously setting and evolving its own goals. That’s not just progress — that’s SKYNET.
@kedidjein
Man your Mona Lisa algorithm is amazing...too!
@somali4154
Google's latest AI advancements might seem impressive, but the pessimist in me sees this as yet another dangerous leap forward for a company that prioritizes profit over people. Handing even more powerful technology to a greedy, ethically hollow corporate giant like Google—one that operates with little real accountability—doesn’t just risk misuse; it threatens to further entrench systems of exploitation, surveillance, and control. Humanity deserves better than having its future dictated by the whims of Silicon Valley’s most indifferent overlords.
@EGarrett01