gerikson

joined 2 years ago
[–] gerikson@awful.systems 11 points 6 months ago

One of the many annoying traits of rationalists is their tendency to backproject classic pieces of literature onto their chosen worldview.

[–] gerikson@awful.systems 9 points 6 months ago (5 children)

Guys according to LW you’re reading Omelas all wrong (just like LeGuin was wrong)

https://www.lesswrong.com/posts/n83HssLfFicx3JnKT/omelas-is-perfectly-misread

[–] gerikson@awful.systems 7 points 6 months ago

Always fun trawling thru comments

Government banning GPUs: absolutely neccessary: https://news.ycombinator.com/item?id=45452400

Government banning ICE vehicles - eh, a step too far: https://news.ycombinator.com/item?id=45440664

[–] gerikson@awful.systems 10 points 6 months ago

Let's not forget the perennial favorite "humans are just stochastic parrots too durr" https://news.ycombinator.com/item?id=45452238

to be scrupulously fair, the submission is flagged, and most of the explicit rat comments are downvoted

[–] gerikson@awful.systems 9 points 6 months ago

This is the dumbest timeline.

[–] gerikson@awful.systems 11 points 6 months ago (10 children)

Some Rat content got shared on HN, and the rats there are surprised and outraged not everyone shares their deathly fear of the AI god:

https://news.ycombinator.com/item?id=45451971

"Stop bringing up Roko's Basilisk!!!" they sputter https://news.ycombinator.com/item?id=45452426

"The usual suspects are very very worried!!!" - https://news.ycombinator.com/item?id=45452348 (username 'reducesuffering checks out!)

``Think for at least 5 seconds before typing.'' - on the subject of pulling the plug on a hostile AI - https://news.ycombinator.com/item?id=45452743

[–] gerikson@awful.systems 4 points 6 months ago

Can't wait to use this as a cudgel when the next lobsters whines that the "vibecoding" tag is disrespectful.

[–] gerikson@awful.systems 9 points 6 months ago

Surprise level: zero

[–] gerikson@awful.systems 6 points 6 months ago

The second screenshot goes to a chart where the Y axis is labelled

Task duration (for humans) where logistic regression of our data predicts the AI has a 50% chance of succeeding

So they're just extrapolating an exponential, not actually measuring it.

[–] gerikson@awful.systems 12 points 6 months ago (9 children)

Check out this epic cope from an Anthropic employee desperately trying to convince himself and others that actually LLMs are getting exponentially better

https://www.julian.ac/blog/2025/09/27/failing-to-understand-the-exponential-again/

Includes screenshots of data where he really really hopes you don't look at the source, and links to AI 2027.

[–] gerikson@awful.systems 10 points 6 months ago (1 children)

He's been on the vibecoding bandwagon for quite some time on lobste.rs.

[–] gerikson@awful.systems 4 points 6 months ago (1 children)

What book is that?

view more: ‹ prev next ›