jamienk

6 ms·

1.
by jamienk 6d ago
ChatGPT accesses my IP address and geo-locates me. Claude code now asks if it can have my browser cookies. Next they will take my address book. They might scan my whole computer. Etc etc. These are pretty low-tech, normal techniques. We can
2.
by jamienk 6d ago
I think OpenAI and Anthropic are slowly feeling the pressure to GET SOME $$ or a plan for some $$ — they need to somehow generate some NETWORK EFFECTS and LOCK-IN. Without that there's no stability: selling ad hoc one-offs is much much
3.
by jamienk 17d ago
There are misplaced verbal emphasis and non-emphasis. And unfortunately these happen at the crucial moments. Makes it very hard to follow.
4.
by jamienk 17d ago
This is different from the final output? I thought all traversals of the chain of language are through these mathematical means. You could tokenize the word "good" to resolve to only represent 'the opposite of "bad"
5.
by jamienk 18d ago
Genuine Q about word optimization/token density: If we create a stripped-down vocabulary with greater token density to use less resources and to resolve ambiguities earlier in the semantic process, aren't we creating NEWSPEAK and
6.
by jamienk 25d ago
The writing reminded me of the SAT test reading section. Impossible to focus.
7.
by jamienk 2mo ago
"Right to repair" isn't some kind of little negotiated contract fiddling. A company can't agree to a 5-year right to repair. Right to repair is a normal freedom, like speech, like using everyday objects you buy or make,
8.
by jamienk 5mo ago
A few things I note: "The family faces immediate FINANCIAL without any legal recourse" WTF? That's not just a flinch, it's some sort of violent tick. The list of "slurs" very conspicuously doesn't include
9.
by jamienk 1y ago
How does this specifically work? Wouldn't any decision about what training data to use be part of a "technique" in this sense? When Stable Diffusion didn't train on porn. OTOH if the majority of your data is "bad&qu