kfue

[kfue @kfue_ai]

@ChillGolfStock @vicmackey24 @ClayTravis What model though

[kfue @kfue_ai]

@menhguin Inequality will increase but everyone will be better off

[kfue @kfue_ai]

@chamath It’s incredible how misleading this is

[kfue @kfue_ai]

You’re quality of life will be SO much better every single year of your life if you’re healthy. If you live unhealthy at 60 you’re going to feel like shit.

If you live healthy at 70 you’ll feel better than the 60 year old.

I’ve always always hated this critique. Being healthy isn’t just about not dying

[kfue @kfue_ai]

@Everlier @OpenAI That is not what that means lol

Training…

[kfue @kfue_ai]

@favelaoverlord If they were truly pricing it in the stock market would be at least a few x higher.

This is what he means. If this happens the amount of wealth creation will be unbelievable.

[kfue @kfue_ai]

@imsodonexD99 @QuietlyRebel @kierstenjonkman @LilahRPGtt An AI can scour 200 sites and find a better answer than I could if I spent 8 hours on a random research curiosity.

It’s not destroying our planet that’s just baseless propaganda

[kfue @kfue_ai]

@arr_kni @benglickenhaus LLMs can do math without tools huh?

Tools enhance their limit but they absolutely can

[kfue @kfue_ai]

@joaomendoncaaaa RLVR? Does training breakthroughs not count?

That’s kinda ridiculous to say

[kfue @kfue_ai]

@intoverflow @halvarflake And the goalposts continue to shift

[kfue @kfue_ai]

@ZootyZootZoot @RichardHanania I don’t think you’ve been paying attention to what the DSA (democratic socialists of America) want if you think no one wants real socialism.

They insane

[kfue @kfue_ai]

@Drachs1978 This is absolutely not true

It’s the way these are trained is why they got good so fast

[kfue @kfue_ai]

@thinkpiecetribe You can also chose not to use a computer to write

Or you could chose to turn a 1h car ride into a full days journey on foot.

But you don’t cause we have technology to do things better now. Use it

[kfue @kfue_ai]

I mean, hasn’t the history of this tech taught us anything by now?

2023: “LLMs will never be good at math”
2024: o1 released. It can start doing some high school competition problems
2025: first “simple” novel proofs solved by LLMs
2026: decades long problems start dropping.

2027 seems prewritten at this point

[kfue @kfue_ai]

@DCinvestor A minority of these math breakthroughs will turn out in to be key in a real world breakthrough

[kfue @kfue_ai]

@Altimor @robbensinger This acts like Yudkowsky’s omniscient. He’s not

Its a very scary thing but the future is not predetermined nor does yudkowsky and co know exactly how it will play out

[kfue @kfue_ai]

@blazin201 @AustroLibertari @BobMurphyEcon Is anything that’s not socialism the right?

[kfue @kfue_ai]

@rand_longevity 8% of all humans who have ever lived are alive today

So it’s not that small but I also think about this

[kfue @kfue_ai]

@ns123abc “Free” lol. Yeah if you have $200k worth of gpus laying around

God the propaganda people spread

[kfue @kfue_ai]

@TrustlessState Such a shitty business model that they’re only the faster growing companies by revenue in history