flaviat

joined 3 months ago
sorted by: new top controversial old
Stubsack: weekly thread for sneers not worth an entire post, week ending 8th June 2025 in c/techtakes@awful.systems
ChatGPT o3 found a Linux Kernel vulnerability. "The future" has an 8% success rate, and a 28% chance of false positives. in c/techtakes@awful.systems
[–] flaviat@awful.systems 24 points 6 days ago

Yet another LLM guy claiming it solved a problem when in fact it was already solved, with it being told almost exactly where and what to look for. Cold reading for use-after-frees.

Stubsack: weekly thread for sneers not worth an entire post, week ending 27th April 2025 in c/techtakes@awful.systems
[–] flaviat@awful.systems 6 points 1 month ago* (last edited 1 month ago) (2 children)

Now I'm wondering if an infinite sequence of nested LLMs could achieve AGI. Probably not.

Stubsack: weekly thread for sneers not worth an entire post, week ending 6th April 2025 in c/techtakes@awful.systems
[–] flaviat@awful.systems 7 points 2 months ago (1 children)

So "unintentional moderates" are those that both choose deliberately to believe in centrist political ideology and to participate in groupthink?

Gemini 2.5 "reasoning", no real improvement on river crossings. in c/techtakes@awful.systems
Credulous coverage of AI slop on Wikipedia in c/techtakes@awful.systems
[–] flaviat@awful.systems 7 points 2 months ago

So perhaps one alternative way to estimate their quality is to check the number of citations, many have more than 100 citations, which is a sign of quality

Andrew Wakefield's 1998 paper has 457 citations on PubMed