blamestross

10 ms·

91.
by blamestross 1y ago
The biggest implications of LLM and programming is this: LLMs are autoencoders for sequences. If an LLM can write the code, the entropy value of that code is low. We know that already, most human communication is low entropy, but the LLMs b
92.
by blamestross 1y ago
Nobody likes the idea that this is only "economical superior AI". Not as good as humans, but a LOT cheaper. The "It will just get better" is bubble baiting the investors. The tech companies learned from the past and they
93.
by blamestross 1y ago
I have such a pair for running longer distances. This scan was part of choosing them. They hurt a lot to get used to (Mostly barefoot for 8 years makes for a hard transition) but now they are good (necessary even) for jogging and distance r
94.
by blamestross 1y ago
Everyone who has said the phrase "Spooky action at a distance" has been proven wrong :)
95.
by blamestross 1y ago
Cool time to show off my gait/foot scan https://my.volumental.com/en/fleet_feet_rics/5128511c-cd88-4... Shoes are evil. Barefoot all the things. I think I might be a bit autistic...
96.
by blamestross 1y ago
Not if it is the authors of the paper filing for the patent. Otherwise people would never publish papers.
97.
by blamestross 1y ago
Its almost like the bottleneck of software engineering "at scale" is human communication not "programming speed"
98.
by blamestross 1y ago
You are not alone! I used programming flow state as pain management for a decade. Its one of those "paid for your mental disorder" situations that are a lot more common than people realize.
99.
by blamestross 1y ago
Its always funny seeing AI safety discussed and pointing at it as "that is just the problems with capitalism in microcosm"
100.
by blamestross 1y ago
You live in a world with MANY AGIs. Collective and Swarm Superintelligence isn't a new thing at all. We call them companies, governments, organizations and churches. They just (mostly) run on meat and memes. The only recent change is t
101.
by blamestross 1y ago
> People would have the same expectations as now If that premise was true, why would misleading advertisement be the norm right now? Why bother? Changing it to a subscription WOULD change perception. Most people don't understand the
102.
by blamestross 1y ago
Ross talks about disclaimers and honesty in advertising being reasonable solutions all the time. He kinda is the "Movement".
103.
by blamestross 1y ago
Right the problem is that they didn't do that. 3 months notice isn't the same as "making it clear to the customer at purchase time how long this product might reasonably be supported" I'd certainly encourage a more
104.
by blamestross 1y ago
Its been one guy and a discord working nonstop for over a year. Not exactly the situation you imagine it. Its also been overwhelming ignored and not-viable in most countries. The US for example is a lost cause. If you think this is things w
105.
by blamestross 1y ago
You could also comply by providing a self hostable server, or even source code that might be convertible to a self hostable server. Requiring an "end of life plan" in the ToS would be a start.
106.
by blamestross 1y ago
You are right, i should have added "like I do" to that statement. Its about the fact i shouldn't be using my internal model as the basis for my expectations of their behavior. I need to suppress that instinct and learn a more
107.
by blamestross 1y ago
The cultural fixation on "Consciousness" has become deeply frustrating to me. We are naturally "animistic" and personifying. The twist in structure allowing mirror neurons being able to re-use the hardware for thinking t
108.
by blamestross 1y ago
My bet is that Dr. Thad Starner's dataset is backing this: https://sites.cc.gatech.edu/home/thad/031_research.htm https://www.popsign.org/ is one of his projects and he has worked with them b
109.
by blamestross 1y ago
Memes are well labeled. A semantic compression of an entire sentiment. The pinnacle of communication by compressed experiential and cultural context. And humans like compressed communication, makes them feel warm fuzzy and agentic when they
110.
by blamestross 1y ago
The nifty feature of intelligence is that you get to examine what you value. Critical peer comments (so far) don't seem to address the "naturalistic fallacy" part, or even lean into it. As far as I am tell, life can be maxima
111.
by blamestross 1y ago
> Which means it would need to sending and receiving the entire state of the CRDT with each change. > If the friend is online then sending operations is possible, because they can be decrypted and merged. Or the user's client can
112.
by blamestross 1y ago
I feel like Wave was "we figured out CRDTs lets go nuts with them" Many of the features that were revolutionary did get baked into docs over the years. I don't think people realize just how similar google docs has become to w
113.
by blamestross 1y ago
Looks a lot like the game theory concept of the "level" of a player. (how many levels of metacognition in others can it practically take into account)
114.
by blamestross 1y ago
Cosmic Background Radiation distribution could be that information. The distribution of mass hitting the event horizon then bouncing
115.
by blamestross 1y ago
*aparent mass goes up Things don't get more mass, they just take more energy to accelerate which looks a lot like more mass. It doesn't imply for example, a high speed mass would cause more gravitational attraction than a slow one
116.
by blamestross 1y ago
The problem is that: - We have a parent universe we will never be able to observe. could be a true statement. The "infinite sequence" part is just a likely implication, it isn't necessarily true. We would need information we
117.
by blamestross 1y ago
Now if only i could afford to spend a few years working on something like this. Sadly there is a mortgage to pay and investment banks who want returns this quarter, not in 3 years.
118.
by blamestross 1y ago
> Does the process matter in all cases? So there are 2 dimensions being conflated here: "Does how the reasoning work matter in all cases" Pretty Obviously no, but it may matter in some of them. We also don't really underst
119.
by blamestross 1y ago
A computer isn't a math capable machine. > Perhaps it is more important to know the limitations of tools rather than dismiss their utility entirely due to the existence of limitations. Well, yes. And "reasoning" is only so
120.
by blamestross 1y ago
I'm kind of upset to see systematically "Alignment" and "AI Safety" co-opted for "undesirable business outcomes". These are existential problems, not mild profit blockers. Its almost like the goals of hum

More ›