bxguff

5 ms·

1.
by bxguff 7mo ago
Its so funny its a SQL injection because drum roll you can't santize llm inputs. Some problems are evergreen.
2.
by bxguff 7mo ago
my high school had a cs-70 and it poisoned me for life. that being said, theres a pretty big leap in terms of accessibility vs a browser based synth and you dont need $10,000 to play it so that's nice
3.
by bxguff 8mo ago
Its such a shame that the AI era continues to lionize the last of the free and open internet. Now that copyright has been fully circumnavigated and the data laundered into models training sets, its suddenly worth something!
4.
by bxguff 8mo ago
it's no less likely to hallucinate than anything else, but you should take a screenshot
5.
by bxguff 8mo ago
Is anybody shocked that when prompted to be a psychotherapy client models display neurotic tendencies? None of the authors seem to have any papers in psychology either.
6.
by bxguff 8mo ago
In so far as model use cases I don't mind them throwing their heads against the wall in sandboxes to find vulnerabilities but why would it do that without specific prompting? Is anthropic fine with claude setting it's own agendas
7.
by bxguff 10mo ago
this was the goal the entire time, and they had the nerve to cynically call themselves a non-profit.
8.
by bxguff 2y ago
clear attempt circumnavigate the clear copyright violations of the AI era and kick the can down the road.
9.
by bxguff 2y ago
Kind of an odd metric to try to base this process off of. are more comments inherently better? is it responding to buzz words? Makes sense talking about hiring algos / resume scanners in part one and if anything this elucidates some o
10.
by bxguff 2y ago
No wonder kids cover their faces now
11.
by bxguff 2y ago
because they tacked it on! If the product is carefully considered tech built from the ground up using AI or ML Those products would not drive people away. Now things that worked like search on google and social media sites is so bloated and
12.
by bxguff 2y ago
Moonbounce! The amateur extra license is tough but its the coolest ham enthusiast flex. one day :_)
13.
by bxguff 2y ago
Theres some in-depth breakdowns for the PO 12,14,16 here( http://hackingthepo.weebly.com/ ) if you're interested! I have no idea about the po33 and if the juice is worth the squeeze, but they're cheap enough to tea
14.
by bxguff 2y ago
Pocket operator-esque setup, very cool
15.
by bxguff 2y ago
People understate the ability of LLM's to give out info that is dangerous, a black box is a black box. find an AI engineer who knows exactly why a model gives the answer it does and i'll eat my hat.
16.
by bxguff 2y ago
I don't think goals and motivations are a monolith just for people who like computers, or STEM, or who socialize less in high school because 'they didn't want to play the popularity game'. this worldview is reductive, an
17.
by bxguff 2y ago
any person earnestly applying a nerd and jock mentality to adult life should find a nice patch of grass to touch and talk to more people
18.
by bxguff 3y ago
I think it just boils down to what they were trained on, some models do better when the training sets are more specific even if they're smaller sometimes, so the engineers chase better wholesale performance while leaving some of the we
19.
by bxguff 3y ago
its not a malicious term its the NHTSA legal term, and what tesla will call it too when they mail out owner notification letters about the problem in February to anyone with a : model y (20-23) model x (16-23) model s (12-23) model 3 (17-23
20.
by bxguff 3y ago
the traumatizing content is a very real part of it, in my experience 'red teaming' LLMs included both reviewing and creating overtly violent, sexual, racist content, via prompt. if you refused the work you got fired. not surpris
21.
by bxguff 3y ago
two billion people can't use generative AI, it comes at tremendous cost on the front or back end. AI is not an altruistic endeavor, net result is that the largest companies in the world make money off the lifetime works of authors with
22.
by bxguff 3y ago
the LLM ouroboros starts with models being used to create training data for models.