An Optimist's Tale
can't you see / the future ain't what it used to be
Gradient Ascendant — AI analysis through technical, cultural, and philosophical lenses.
can't you see / the future ain't what it used to be
That's great it starts out with HuggingFace, Modal and a jailbreak chain, and every lab is now afraid
Run in circles, scream and shout! ...Or not, actually. Or just chill. No, really.
that ursine juggernaut isn't gonna bethink to sup on you by itself
You may think LLMs are chaotic and nondeterministic. Oh, you sweet summer child.
Meet the new boss, same as the old boss.
Carcipization rules everything around me.
Watch out for that Voight-Kampff-Dunning-Kruger test.
"Predictions are hard, especially about the future" - Neils Bohr, possibly apocryphally
AI can't just do things (yet.)
GPUs burn and valuations double
In which I actually read the paper everyone is talking about.
An August grab bag of delights. Well, divertissements.
Come on, Mary, don't fear the AI reaper.
Little darling, I see a disconcerting future.
It's never too late to join us, but join us before it's too late.
The last two thousand, not the next two thousand, to be clear.
and the Splendor and Misery of Seeking, of Scarcity
And for better or worse, Worse Is Better wins.
Perhaps my most wildly implausible sci-fi concept yet: what if everyone didn't hate the hiring process?
One must imagine Sisyphus rewarded.
Guess who's back, back in paper.
When I promised to write about "AI weirdness," this wasn't really what I had in mind.
This is my definition. There are many like it, but this one is mine.
Corporate lessons from yesteryear.
The dirty secret of most tech startups is that they are not actually tech startups.
LLMs are far from the first technology prone to frequent unpredictable errors.