zota
@zota.bsky.social
Technology therapist, gonzo designer, grey-hat historian.
Los Angeles.
Los Angeles.
Or just putting it in an index fund, forgetting it exists, and compounding incomprehensible wealth smdh
December 11, 2025 at 7:51 PM
Or just putting it in an index fund, forgetting it exists, and compounding incomprehensible wealth smdh
Reposted by zota
becoming a really good writer is an extensive and lengthy process that also requires reading very widely
the thing about inventing LLM tech that relies upon natural language to function is that it becomes even more important that the humans doing the prompting are good at using natural language
the thing about inventing LLM tech that relies upon natural language to function is that it becomes even more important that the humans doing the prompting are good at using natural language
December 11, 2025 at 6:07 PM
becoming a really good writer is an extensive and lengthy process that also requires reading very widely
the thing about inventing LLM tech that relies upon natural language to function is that it becomes even more important that the humans doing the prompting are good at using natural language
the thing about inventing LLM tech that relies upon natural language to function is that it becomes even more important that the humans doing the prompting are good at using natural language
"Wouldn't it be nice to get a massage?"
December 10, 2025 at 10:36 PM
"Wouldn't it be nice to get a massage?"
If you were a real Texan, you'd be honest about the pieces of shit who rule it with lies and corruption.
Maybe you're a Canadian immigrant like your Senator?
Maybe you're a Canadian immigrant like your Senator?
December 10, 2025 at 4:46 PM
If you were a real Texan, you'd be honest about the pieces of shit who rule it with lies and corruption.
Maybe you're a Canadian immigrant like your Senator?
Maybe you're a Canadian immigrant like your Senator?
If you were honest about the state where I was born, you wouldn't lie about "liberal enclaves"—aka the cities where most people live—being allowed to "do what they want." The maga oligarchs made that illegal.
They also gerrymandered those "enclaves" to take away the right to vote.
They also gerrymandered those "enclaves" to take away the right to vote.
December 10, 2025 at 4:44 PM
If you were honest about the state where I was born, you wouldn't lie about "liberal enclaves"—aka the cities where most people live—being allowed to "do what they want." The maga oligarchs made that illegal.
They also gerrymandered those "enclaves" to take away the right to vote.
They also gerrymandered those "enclaves" to take away the right to vote.
Bro. You're a Canadian talking about how you have no pity for the American victims of fascism.
Don't you have some racists to deal with at home?
Don't you have some racists to deal with at home?
December 9, 2025 at 3:58 AM
Bro. You're a Canadian talking about how you have no pity for the American victims of fascism.
Don't you have some racists to deal with at home?
Don't you have some racists to deal with at home?
It feels good to think about those people getting what they deserve, doesn't it? Finally being in power over those scum? Those ignorant subhumans?
That's the entire emotional foundation of fascism.
That's the entire emotional foundation of fascism.
December 9, 2025 at 3:48 AM
It feels good to think about those people getting what they deserve, doesn't it? Finally being in power over those scum? Those ignorant subhumans?
That's the entire emotional foundation of fascism.
That's the entire emotional foundation of fascism.
My goal was to understand how you deal with contradiction and ambiguity in the data, so I could better frame you as a source for information and opinion.
cmr.berkeley.edu/2025/10/seve...
I've gotten my answer, so yes, it is best we move on
cmr.berkeley.edu/2025/10/seve...
I've gotten my answer, so yes, it is best we move on
Seven Myths about AI and Productivity: What the Evidence Really Says
Meta-analytic evidence finds no robust relationship between AI adoption and aggregate productivity gain.
cmr.berkeley.edu
December 9, 2025 at 3:22 AM
My goal was to understand how you deal with contradiction and ambiguity in the data, so I could better frame you as a source for information and opinion.
cmr.berkeley.edu/2025/10/seve...
I've gotten my answer, so yes, it is best we move on
cmr.berkeley.edu/2025/10/seve...
I've gotten my answer, so yes, it is best we move on
If the "reliable" source you cited as evidence that AI has not plateaued is in fact completely unreliable on the question of how useful AI actually is, making a firm statement of truth on the issue seems more like a statement of belief.
December 9, 2025 at 3:09 AM
If the "reliable" source you cited as evidence that AI has not plateaued is in fact completely unreliable on the question of how useful AI actually is, making a firm statement of truth on the issue seems more like a statement of belief.
"actual productivity uplift from using AI agents lags their perceived benchmark capabilities, with our early-2025 uplift study finding no (or very limited) productivity benefits from AI usage for experienced open-source developers even for short tasks"
Thanks for the link. That is clarifying.
Thanks for the link. That is clarifying.
December 9, 2025 at 3:01 AM
"actual productivity uplift from using AI agents lags their perceived benchmark capabilities, with our early-2025 uplift study finding no (or very limited) productivity benefits from AI usage for experienced open-source developers even for short tasks"
Thanks for the link. That is clarifying.
Thanks for the link. That is clarifying.
The promise of AI—as demonstrated in the chart itself—is that soon line will go up infinity times, to the moon.
If the rate of improvement has declined by 80%, is it unfair to call that a relative plateau?
If the rate of improvement has declined by 80%, is it unfair to call that a relative plateau?
December 9, 2025 at 2:53 AM
The promise of AI—as demonstrated in the chart itself—is that soon line will go up infinity times, to the moon.
If the rate of improvement has declined by 80%, is it unfair to call that a relative plateau?
If the rate of improvement has declined by 80%, is it unfair to call that a relative plateau?
If you want to have an honest discussion, that's a very weird chart to use. It asks me to accept a *lot* of assumptions as given (including, the source).
But taking it entirely on its own terms: "early" models were line go up 100% a year. Current models are line go up at maybe 20%?
But taking it entirely on its own terms: "early" models were line go up 100% a year. Current models are line go up at maybe 20%?
December 9, 2025 at 2:50 AM
If you want to have an honest discussion, that's a very weird chart to use. It asks me to accept a *lot* of assumptions as given (including, the source).
But taking it entirely on its own terms: "early" models were line go up 100% a year. Current models are line go up at maybe 20%?
But taking it entirely on its own terms: "early" models were line go up 100% a year. Current models are line go up at maybe 20%?
Have there been major articles proving AI hasn't plateaued?
December 9, 2025 at 2:30 AM
Have there been major articles proving AI hasn't plateaued?
Ah.
Leonard Rodberg—Research Director, Nuclear New York—is threatening California with blackouts if we don't buy reactors from Bill Gates.
Yeah, I can see why you'd want to downplay that.
Leonard Rodberg—Research Director, Nuclear New York—is threatening California with blackouts if we don't buy reactors from Bill Gates.
Yeah, I can see why you'd want to downplay that.
December 9, 2025 at 2:26 AM
Ah.
Leonard Rodberg—Research Director, Nuclear New York—is threatening California with blackouts if we don't buy reactors from Bill Gates.
Yeah, I can see why you'd want to downplay that.
Leonard Rodberg—Research Director, Nuclear New York—is threatening California with blackouts if we don't buy reactors from Bill Gates.
Yeah, I can see why you'd want to downplay that.
mass murder of her career?
December 8, 2025 at 8:27 PM
mass murder of her career?
The whole thing is 125 lbs suddenly going dead weight. The arm alone is probably close to 20 lbs
December 8, 2025 at 3:46 PM
The whole thing is 125 lbs suddenly going dead weight. The arm alone is probably close to 20 lbs