antineurotic

nonpartisan generalist epistemologist

[antineurotic @antineuro]

@ersatz_0001 @segyges This probabilistic reasoning lets people get away with all kinds of bad predictions

Everything becomes unfalsifiable

[antineurotic @antineuro]

So this is obviously wrong but what's interesting to predict is how they're going to respond when this turns out be wrong, and my guess is that they'll call everything 'artisanal'

[will depue @willdepue]

@niloofar_mire @OpenAI dude i hate to break it to you but it’s totally fucking over for the human race and any & all nonartisanal (derives value from being done/created by a human) economic labor. we will soon all watch from the sidelines. it’s magnus carlsen vs stockfish 13 from here on out

[antineurotic @antineuro]

Everyone’s talking about how the superintelligence story is nonsense now and that makes me happy

[RAY @4xCaliban]

I mean the obvious conclusion is that the market does not believe that Anthropic will do any of these things.

[antineurotic @antineuro]

@archiemckenzie_ @pangram Completely antithetical to the rest of the AI industry; superintelligence will vary its writing enough to avoid detection. Pangram's model is extremely pessimistic, but maybe more correct

[antineurotic @antineuro]

Well-defined* instead of verifiable

We don't even clearly know what the objective / nature of all these other domains is, e.g. writing

[antineurotic @antineuro]

This is all downstream of text models being used for everything. Text is not the solution to everything (although clearly it can be used to solve many things like math and code).

And it’s also unclear whether text itself will be solved. It’s been a while with no progress.

[Greg Burnham @GregHBurnham]

@tallinzen From some recent work trying to get AI to automate some of our own work, this sparse human vs dense AI information style is quite something. (Just a graphic, not a slide; but might as well be.)

[antineurotic @antineuro]

Code and math (only some parts) are only solved because they are verifiable domains! There is no reason to think _any_ unverifiable domain (almost everything else) is solvable. But this is the default belief

[antineurotic @antineuro]

It is more hilarious (and common) to see people assume the opposite: superintelligence is certain. That there is no recalcitrance in intelligence growth.

This is a _possible_ (and high payoff, so worth emphasizing) outcome, but very unlikely.

[Vamsi Pingali @vamsiprithamp]

It is hilarious to see people assuming that AI will remain frozen at today’s level forever: https://t.co/PwKb1ZDIeL

[antineurotic @antineuro]

https://t.co/uUHAtCWUxP

[Theo - t3.gg @theo]

Crazy how Large Language Models have gotten good at pretty much everything other than generating text

[antineurotic @antineuro]

Everyone sees it

https://t.co/X4gREeT3tj

[coah @coah80]

bathroom if it was vibecoded:

Poop
Your poop, Sitting there

Shower
Water hits your body, cleaning it.

Sink
Brush your teeth, or wash your face

[antineurotic @antineuro]

https://t.co/qpmIYn8wRV

[forward deployed ccp gf @FangYi11101]

@qquinnsworth One of the most incompetent guys I worked with was top 10 globally in Hearthstone

[antineurotic @antineuro]

People who don’t game all day don’t realize that the people who do game all day are pretty stupid. Intelligence of ‘gamers’ is massively overrated.

[alz @alz_zyd_]

there is a certain personality type which will play Civilization or Factorio 14 hours a day pausing only to sleep and microwave food. this type can be absurdly economically productive if you can take the games away, give him a job, and get him to think of it as a videogame

[antineurotic @antineuro]

Second time recently that rightoids have tried and failed dramatically at statistics. First was "correlation isn't causation" at the Starbucks productivity paper. I expect this to continue.

[Owen Winter @OwenWntr]

@JohnLeFevre It’s on a log scale because that’s the appropriate scale for ratios. Being half as likely to be convicted should be as far from x=0 as as being twice as likely

[antineurotic @antineuro]

@RichardHanania All skepticism gets thrown out the window when discussing people who are considered the best at something

[antineurotic @antineuro]

@guilleflorvs @OpenAI @AnthropicAI please solve writing I’m so tired

[antineurotic @antineuro]

@Leothebayesian @ScottWu46 Math people are too critical of themselves. Any other field would embrace any good way to market itself, even if inaccurate. Math people say things like this.

[antineurotic @antineuro]

@ByrneHobart Even the thought is ridiculous

[antineurotic @antineuro]

@oecolamp Not really

[antineurotic @antineuro]

True. There is life after AI, but this isn’t it.

[BURKOV @burkov]

It's so funny to see people who never used AI to code still think in categories like "OK, but then it will break the production and no one will be able to understand what's going on to fix it!"

They cannot imagine that no one will be needed to understand anything: They will just send a new prompt "This thing produced this wrong value. Find the reason and fix it."