343rwerfd

6 ms·

1.
by 343rwerfd 3mo ago
I think any new model not demonstrably maybe 20-30% over Deepseek v4 capabilities priced over the price per token of Deepseek is almost automatically deprecated as low use model (maybe for Planning).
2.
by 343rwerfd 2y ago
Training an AGI/ASI does not requires the biggest datacenters/massive GPUs, nor it takes years already. Early algorithmic advances and narrow AGI AIs have radically shortened the requirements in hardware and time of training. You
3.
by 343rwerfd 2y ago
You're mentioning only publicly known information. The rumors mentioning radical advances behind closed doors are wild, and then you've suddenly got some stuff like deepseek or phi-4. Rumors mention recursive "self" impr
4.
by 343rwerfd 2y ago
"Why lower the bar?" Because of the chance of misundertanding. Failing at acknowledging artificial general intelligence standing right next to us. An incredible risk to take in alignment. Perfect memory doesn't equal to perfe
5.
by 343rwerfd 2y ago
Deepseek completely changed the game. Cheap to run + cheap to train frontier LLMs are now in the menu for LOTs of organizations. Few would want to pay AI as a Service to Anthropic, OpenAI, Google, or anybody, if they can just pay few millio
6.
by 343rwerfd 2y ago
Since frontier models evolved beyond the very basic stuff from maybe 2020, "LLM can only make predictions of word sequences" only describes a small fraction of the inner processes that the frontier systems use to get to the point
7.
by 343rwerfd 2y ago
I think the concepts underlying the whole LLM technological ecosystem are currently quite new, the best they can do is to use some refurbished familiar language, somewhat aligned with the aproximate (probable?) actual meaning in the the con
8.
by 343rwerfd 2y ago
"probabilistic storytelling engine" It's a bit more complicated thing than that. You most probably could describe it as something capable of exercising the same abilities that humans and other species exercise when they use a
9.
by 343rwerfd 2y ago
In Argentina, this pro-Austrian economics government which has severe limitations in terms of law, regulations, and the heavily destroyed general economy of the country, does not have enough freedom to swiftly change things to "what id
10.
by 343rwerfd 2y ago
a possible lesson to infer from this example of human cognition, would be that LLMs that can't solve the strawberry test could not be automatically less cognitive capable that another intelligent entity (humans by default). An extensio
11.
by 343rwerfd 2y ago
Not necessarily a happy story,though
12.
by 343rwerfd 2y ago
The hidden chain-of-though inside the process, from the official statement about it, I infer / suspect that it uses an unhobbled mode of the model, puts it in this special mode where it can use the whole training, avoiding the intrisic
13.
by 343rwerfd 2y ago
IT salaries began to go down right after AI popped up out of GPT2, showing up not the potential, but the evidence of much improved learning/productivy tool, well beyond the reach of internet search. So beyond, that you can easily can t
14.
by 343rwerfd 2y ago
ASI is the endgame where it is profitable to be in the OpenAI position, or even in the next first 20 market players capable of getting there a bit later. But if ASI isn't achievable finally, the intelectual properties obtained in the w
15.
by 343rwerfd 2y ago
Claude Sonnet 3.5 at least works awesomely, you can just talk to it asking stuff, it will infer your knowledge level from your questions and start answering according to it, proposing a follow up path for further insigths about the subject.
16.
by 343rwerfd 2y ago
The industry always has information before-hand, all those AI capable datacenters aren't being built just because a hunch. It is possible that the next iteration of GPT4 level technology has already ocurred a year ago, august 2023. The
17.
by 343rwerfd 2y ago
Yes, this happens, there's happening some throttling, I've seen questions like this one regarding the same issue across several LLM providers ("works faster, better, solves better at night").
18.
by 343rwerfd 2y ago
Lots and lots words flow about this. For me, it is very simple. LLMs do a complex process, quite analog to human thinking/understanding/speaking/writing, but they're doing all those - most probably - in an alternative wa
19.
by 343rwerfd 2y ago
Moroever, those millions of non-paying clients prompting the models are 24x7x265 working with 100% real world problems are inputing the models with valuable prompts, generating valuable content (originated in real situations, actually disti
20.
by 343rwerfd 2y ago
nor OpenAI or any of the prompt-based AI companies actually "need" the reveneu from the services they sell, the whole point of having a public (free or not), prompt facing the entire planet is just having live humans doing RLHF 24
21.
by 343rwerfd 2y ago
Not really, China is just a step behind, in a year they will be at current US AI state-of-art, without competition, from there they'll have all the GPUs in the world to keep improving their models. Or most probably, US national interes
22.
by 343rwerfd 2y ago
> If an LLM was capable of logical reasoning the prompt interfaces + smartphone apps were (from the beginning), and are ongoing training for the next iteration, they provide massive RLHF for further improvements in already quite RLHFed a
23.
by 343rwerfd 2y ago
> Next-token prediction cannot handle Kuhnian paradigm shifts The training datasets contain and reflect human imperfections, including implicit mistakes (which could not be fully extracted in the refining work because ambiguity), and fai
24.
by 343rwerfd 2y ago
> Conflicts (such as an attempt to kill humanity) have no zero-risk moves I think the author is onto something here, probably correct. Hence, most theories of AGI doomsday relay on AIs deception and asymmetrically deployed actions. Some
25.
by 343rwerfd 2y ago
>there is a relationship between information theory and thermodynamics, and nobody, including no superintelligence, will be able to break it. The way the lesswrong people could know or not, to "break" through information theory
26.
by 343rwerfd 2y ago
> No progress without experiments The "experiment" is us. The prompting interfaces face the entire human population, a sizable number is currently feeding the models with valuable/actionable experiments plus outcomes, re-f
27.
by 343rwerfd 2y ago
>For an AI, a thought experiment and a real experiment are indistinguishable. >As a result, any world model that is learnt through the analysis of text is going to be a very poor approximation of reality. As of recently, and regarding
28.
by 343rwerfd 2y ago
>However it also fails in other areas quite spectacularly: > Anything which requires logic > Anything which requires actual understanding Not really the experience most people is getting at using AIs nowadays. It actually shows a h