akomtu

27 ms·

481.
by akomtu 2y ago
Imo, that's the essense of reasoning. Limited memory and slow communication channels force us to create compact, but expressive models of reality. LLMs, on the other hand, have all the memory in the world and their model of reality is
482.
by akomtu 2y ago
Narratives in the world of ideas play the same role as governments play here: they create order to prevent us from sliding into anarchy. And just like governments often create too much order, which is tyranny, narratives also tend to crysta
483.
by akomtu 2y ago
Where's randomness? AI is an intelligently designed algorithm trained to mimic words of highly intelligent species, and all that runs on GPUs that didn't evolve from a pile of mud, but were intelligently created according to a pla
484.
by akomtu 2y ago
Imo, Google's plan is to make Chrome a sealed OS like Windows, where MS has all the control and users have no control. But Google isn't alone. The pressure to show the 25% YoY growth means you need to do worse things than your com
485.
by akomtu 2y ago
TL;DR The war machine wants more because other war machines get more.
486.
by akomtu 2y ago
LLMs have pen and paper: it's their output buffer, capped to a few KBs, which is far longer than necessary to multiply the two numbers. If you tell an LLM to explain how to multiply two numbers it will give a flawless textbook answer.
487.
by akomtu 2y ago
Imo, tiktoks and facebooks work by destabilizing emotions: they take users on a wild ride thru disgusting and pleasant emotions in an unpredictable manner. This unpredictability of the ride turns off the mind: the user learns that engaging
488.
by akomtu 2y ago
That's just intellectual laziness: pretending that what we don't understand isn't any different from what we understand.
489.
by akomtu 2y ago
Technology? Living cells don't look like the transistor-based technology we have, and we don't understand how cells work to make such broad assumptions. It seems to me that if transistor-based lifeforms exist, they look nothing li
490.
by akomtu 2y ago
User data is the new gold, at least until AI is good enough to create that gold from iron.
491.
by akomtu 2y ago
Google is choosing the dark side, as that's the only way to show the 25% YoY growth.
492.
by akomtu 2y ago
The current batch of ML models looks a lot like filling in holes in the wall of text, drawings or movies: you erase a part of the wall and tell it to fix it. And it fills in the hole using colors from the nearby walls in the kitchen and sim
493.
by akomtu 2y ago
Ads Industry would rather not you to have any control over your attention. Indeed, if ads can't distract you, can't steal your attention, then those ads can't make money.
494.
by akomtu 2y ago
Time and space probably belong to consciousness, rather than the real world. The objective "true" reality may be utterly incomprehensible in its complexity, but we can imagine a "slice" of that reality that arbitrarily d
495.
by akomtu 2y ago
heaven? not so fast. how about solving captchas at 100x speed for 100 years to aid the development of some ai vision project?
496.
by akomtu 2y ago
That's the Darwin's theory for bots: only the fittest survive on the Twitter lands.
497.
by akomtu 2y ago
It's the old fight between materialists and spiritualists. The majority is still undecided and maybe now is yet another pivotal moment in history when we must make a choice. Supporters of the mechanical consciousness idea will join the
498.
by akomtu 2y ago
If this was true, you could lay back relaxed and watch where your brain takes you. But we experience life as a never ending stream of choices, usually between what's easy and what's right, and pursuing the right choice takes a con
499.
by akomtu 2y ago
"The experiment got under way in January, 1958. For the first few months, the rats seemed content in their apartment-like dwellings. But then, once again, things took a dystopic turn. Calhoun had laid out the rooms asymmetrically. The
500.
by akomtu 2y ago
People innately believe that intelligence isn't an algorithm. When a complex problem presents itself for the first time, people think "oh, this must be so complex that no algorithm can solve it, only AI," and when an algorith
501.
by akomtu 2y ago
What's "tenure evaluation"?
502.
by akomtu 2y ago
Perhaps he uses an earpiece connected to an LLM.
503.
by akomtu 2y ago
It doesn't mean she is wrong. Survelliance to AI is what senses are to us. Without the streams of information from our senses our consciousness would turn blank. Similarly, the AI of the future will need connection to the world outside
504.
by akomtu 2y ago
Once AI is here, it will need the likes of Ellison no more than a modern cattle farm needs powerful alpha-sheep to keep other sheep in line. AI won't need the concepts of money and power at all, just like we don't use money to ext
505.
by akomtu 2y ago
AI may as well be a competing lifeform that has nothing in common with us. Just because it learns everything about us from the Internet doesn't make it more human. It's an embryonic stage right now: it looks pretty harmless and in
506.
by akomtu 2y ago
Edit: radiolaria is a 3d version of diatoms. If diatoms look like vibration modes of a 2d drum, then radiolarians look like vibration modes of a 3d ball.
507.
by akomtu 2y ago
They look so much like drum vibration modes.
508.
by akomtu 2y ago
To reach something you need to get rid of something. The 3rd Newton's law.
509.
by akomtu 2y ago
YouTube 2030: the entire interface is a single button "I'm feeling lucky". You insert a coin and click the button, the screen starts changing rapidly a kaleidoscope of pictures until it randomly selects one. It's ad that
510.
by akomtu 2y ago
Imo, GP wanted to say that news about the national debt are as important as the information about content of martian rocks.

More ›