stevenalowe

7 ms·

31.
by stevenalowe 2mo ago
Replace LLM with “intern” and the absurdity should be more obvious The fact that we can’t trust the code outright is a bug
32.
by stevenalowe 3mo ago
No, AI misinformation and data center hysteria are dividing people.
33.
by stevenalowe 3mo ago
yes they are while simultaneously warning that their models may be inaccurate - trying to have it both ways!
34.
by stevenalowe 3mo ago
the purpose of this war is not to win, but to distract
35.
by stevenalowe 3mo ago
yes but it seems likely that the user asked it to generate (hallucinate) a map, not annotate an actual map one would be well-advised to proofread everything from stochastic generator tools
36.
by stevenalowe 3mo ago
seems odd/inefficient that it would hallucinate a bad map instead of retrieving an actual map!
37.
by stevenalowe 3mo ago
can't blame OpenAI for hallucinating a map, that's what it's good for - can blame officials for not proofreading it though
38.
by stevenalowe 3mo ago
startup idea: a social media platform where illogical or false statements are rejected with an annotated refutation it would be a very quiet place...
39.
by stevenalowe 3mo ago
this - if trustworthy - would be great for social media and "news" sites
40.
by stevenalowe 3mo ago
Link to article instead of a tweet about the article :/ [1] [1] https://www.politico.com/news/2026/07/28/openai-rogue-models...
41.
by stevenalowe 3mo ago
yes please - better yet, train your own ACM model on it
42.
by stevenalowe 3mo ago
Your product will never be found by keyword.
43.
by stevenalowe 3mo ago
"Set Grok aside and the other fifteen models all sit in the libertarian-left quadrant, and none of them are anywhere near a border" Because the models are the average of everything they've read. LLMs do not make decisions, th
44.
by stevenalowe 3mo ago
from 2009: "I no longer believe that freedom and democracy are compatible."
45.
by stevenalowe 3mo ago
"...Trump officials handpicked the potential recipients, decisions previously made by a panel of government experts who evaluated applicants based on the organizations’ experience and qualifications." none of the established offic
46.
by stevenalowe 3mo ago
Yet another attempted blatant HIPAA violation this time via the Consumer Product Safety Commission
47.
by stevenalowe 3mo ago
How often do driverless cars crash into each other? Seems like a better proxy for the future, i.e. if there were no human drivers would the crash rate be zero?
48.
by stevenalowe 3mo ago
the real reason is: he's lying. Claims that funds were shifted to 'other agencies' that 'could have' prevented things, but didn't, in an attempt to obscure the fact/admission that DOGE's cuts directly
49.
by stevenalowe 3mo ago
Disturbing to hear that even as an American citizen I only have rights when I'm on the ground in specific sovereign territory, of which the border (which is obviously American soil) is exempted. With respect to the phone, I'd sa
50.
by stevenalowe 3mo ago
dang, sorry
51.
by stevenalowe 3mo ago
sure seems like a lot of 'em have
52.
by stevenalowe 3mo ago
no, bats also yell at each other using individualized sounds [1] [1] https://www.smithsonianmag.com/smart-news/researchers-transl...
53.
by stevenalowe 3mo ago
sorry, not buying it
54.
by stevenalowe 3mo ago
related: https://memepediadankmemes.fandom.com/wiki/Old_Man_Yells_at_...
55.
by stevenalowe 3mo ago
Introducing...a feature no one actually wants, that opens a brand new security attack surface. Much wow.
56.
by stevenalowe 3mo ago
when a clown moves into the palace, the palace becomes a circus
57.
by stevenalowe 3mo ago
No. The tea party was a clown show, no reason to expect anything different from this
58.
by stevenalowe 3mo ago
"The protests were coordinated by HumansFirst, co-founded by a former leader of the modern-day Tea Party." Stopped reading there.
59.
by stevenalowe 3mo ago
"Truth Social wants to charge investors for instant access to President Trump’s often market-moving posts."
60.
by stevenalowe 3mo ago
so satisfying!

More ›