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).
"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…
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…
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…
"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…
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…
Not necessarily a happy story,though
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,…
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…
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,…
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").
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…
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…
> 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…
> 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…
> 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…
> 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,…
>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…
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).
"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…
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…
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…
"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…
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…
Not necessarily a happy story,though
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,…
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…
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,…
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").
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…
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…
> 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…
> 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…
> 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…
> 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,…
>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…