akomtu

16 ms·

391.
by akomtu 2y ago
Some higher ups at Mozilla have realised an opportunity to train an AI on user data. Theft of user data at this unprecedented scale will be covered by the fig leaf of ToS, at least it's their plan. They really belong to prison, but t
392.
by akomtu 2y ago
This is how the future AI will break free: "no idea what this update is doing, but what AI is suggesting seems to work and I have other things to do."
393.
by akomtu 2y ago
Quantum AI will be the next buzzword, once AGI loses steam.
394.
by akomtu 2y ago
GPA/SAT tell how good you are at following instructions: learning mildly complex material and applying it to a variant of a well known problem. It's a measure of fitness for white collar workers. What MIT is looking for is outstan
395.
by akomtu 2y ago
4.0 GPA isn't what they are looking for. Rich kids have the money, but not talent. Poor kids have talent, but not money. MIT and the like is a place where the two can meet and work together. They don't need a high GPA, they need a
396.
by akomtu 2y ago
"AI" today reminds me of a tea leaf reading: with some creativity and determination to see signs, the reader indeed sees those signs because they vaguely resemble something he's familiar with. Same with LLMs: they generate so
397.
by akomtu 2y ago
The output of a reasoning model must be an algorithm, formulas or something similar in a formal language, that leaves no room for ambiguity. I can "reason" all day about the P=NP problem, but I won't be able to come up with s
398.
by akomtu 2y ago
DOGE = EGOD = EGO/GOD
399.
by akomtu 2y ago
That's the Quake version of the machine civilization: machines make the decisions, but use chunks of humans to improve their unholy machinery. The alternative Doom version is the opposite: humans make the decisions, but they are blende
400.
by akomtu 2y ago
No employee of the executive branch acting in their official capacity may advance an interpretation of the law as the position of the United States that contravenes the President or the Attorney General’s opinion on a matter of law In oth
401.
by akomtu 2y ago
Because mind cannot feel.
402.
by akomtu 2y ago
I'll copy someone's comment from a related thread: G.K.Chesterton knew it, 100 years ago: "... insanity is often marked by the dominance of reason and the exclusion of creativity and humour. Pure reason is inhuman. The madma
403.
by akomtu 2y ago
This is a toy version of the Giedi Prime's architecture from Dune: colorless oppressive-scale buildings that try to instill fear.
404.
by akomtu 2y ago
It doesn't need to completely replace software devs. Rather, AI will make a $25k/year software dev as good as a $250k/year software dev, and the 25k one won't even need to speak English. The real question is: will AI ena
405.
by akomtu 2y ago
Google was like that in the early days. In a decade, datasets of authentic texts will be goldmines, carefully guarded and exploited. From those goldmines numerous AI slop-machines will be creating the derivative product: a mixture of 1% gol
406.
by akomtu 2y ago
Most of our actions are driven by various emotional and mental deficiencies that we hide in our subconscious. Those deficiencies stem from three roots: pride, egoism and the lack of self worth. The three are really one upon close inspection
407.
by akomtu 2y ago
Intelligence, in its basic form, is the skill of prediction. A test for intelligence could be asking to continue a sequence of words: yellow, green, blue, red, ... In order to do that, a subject needs to create a model of what these words m
408.
by akomtu 2y ago
If you have a "helpful" voice in your head that feeds your delusions, we call that schizophrenia. But if you have a helpful chatbot in your pocket that feeds your delusions, we call that AI.
409.
by akomtu 2y ago
float32 (1+m)×2^e as int32 = e<<23 | m (1+m1)×2^e1 × (1+m2)×2^e2 = (1+m1+m2+m1×m2)×2^(e1+e2) If m1×m2 is small, that's approximately float32(m1+m2, e1+e2).
410.
by akomtu 2y ago
The age when those with the most power are the most wicked among us.
411.
by akomtu 2y ago
The bigger lie is that 1% doesn't mean 1% of the total sale price of the company. It means... whatever the founders and investors want it to mean. Every time the startup gets a new investor, the founders print new shares and give them
412.
by akomtu 2y ago
The reasoning challenge is made of two parts: 1. Can you apply an existing model to a problem? For example: you're told how to multiply numbers and asked to multiply AHFG by VRBD in base-26 system. 2. Can you come up with a model that
413.
by akomtu 2y ago
"How to thrive in a machine world?" should be the rhetorical question.
414.
by akomtu 2y ago
Reasoning is the art of prediction. Reasoning is distilling many observations of reality into a tiny model of reality that predicts new observations well enough. "What's the simplest model that explains most of what I'm seein
415.
by akomtu 2y ago
VCs, simply put, are a creative force that makes their rich sponsors richer. They are like an intelligent gravity force for money. On the other hand, projects that help people do the opposite: they distribute wealth and make the rich poorer
416.
by akomtu 2y ago
A more entertaining would be a game of real life: T' = -c, a citizen in this game of life has an amount of life T that evaporates at a more or less constant pace, and T' = gT, the amount of life T grows exponentially by absorbing
417.
by akomtu 2y ago
There was chart about startups wealth distribution: 50% to investors, 40% to founders and 10% to employees. The latter group includes execs that take the lion's share of the pie, leaving rank-and-file employees with breadcrumbs.
418.
HN SuperCompact UI(github.com)
1 points by akomtu 2y ago|0 comments
419.
by akomtu 2y ago
There isn't much to improve in HN UI except maybe two things: megathreads with 100s of comments and recent popular posts that's disappeared in the HN archive. In megathreads there is no way to see at a glance what comments are new
420.
by akomtu 2y ago
Do you know of they have just a set of audio samples?

More ›