pu_pe

6 ms·

61.
by pu_pe 4mo ago
Europe is in the worst spot right now, because even if open source is the future, there is not enough European-owned datacenters even for inference. Not to mention that China could pull the rug on these models at any moment just like the US
62.
by pu_pe 4mo ago
Without access to leading models, I think open source LLM development will also slow down. I'm not sure which portion of their success right now is due to RLAF and distillation but it's certainly not zero.
63.
by pu_pe 4mo ago
Argentina's interest rate is ~30% right now. So I would argue both the investors and the Argentinian government are aware of how risky these loans are.
64.
by pu_pe 4mo ago
The article doesn't really mention this, but it would be far more realistic, constructive and pleasant to focus on how we can all reduce our work hours with AI rather than straight out mass unemployment. In fact I think political movem
65.
by pu_pe 4mo ago
It's true that no one is trying to one shot anything serious right now, but it's still an important metric. Claude Code and Opus really took off when they improved the harnessing enough that it would self-correct many of its mista
66.
by pu_pe 4mo ago
The fact this Midjourney delusional device is taken even slightly seriously is a major sign that we are in a bubble. A lot of people are pointing to Anthropic and OpenAI's investments into datacenters as problematic, but I think it&#x2
67.
by pu_pe 4mo ago
So in at least two fields in which the US still leads (biomedical research, AI), government action is actively harming their ability to compete because of ideological or political reasons.
68.
by pu_pe 4mo ago
You can run it, but you can't train it. While this type of toy model could actually be trained in Swiss equipment, a state-of-the-art LLM probably could not. My charitable reading of GP's point is that the bottleneck for true comp
69.
by pu_pe 4mo ago
It's not like their 10x revenue growth year-to-year was based on expanding their customer base 10x though. The vast majority of weekly users are not paying, so they are actually a cost right now, but could be converted into paying user
70.
by pu_pe 4mo ago
The Amodei quote they use is illuminating: "The government should have the power to block or deter deployment of the model if it is determined, in light of third-party assessment, to present unacceptable risks. This power must be scope
71.
by pu_pe 4mo ago
Do you have a source for this?
72.
by pu_pe 4mo ago
It stinks to high heaven, especially considering how over-the-top security protocols were introduced with Fable. The US government is asserting its influence on the economy and showing Anthropic that their IPO will depend on bending the kne
73.
by pu_pe 4mo ago
The isochrone maps are quite beautiful [1], and go beyond the scope and refinement of some earlier human attempts I could find [2][3][4]. [1] https://isochronic-passage-chart.netlify.app/ [2] https://mapitout.wel
74.
by pu_pe 4mo ago
I never said an organisms behavior outside of reproduction does not have an impact on natural selection.
75.
by pu_pe 4mo ago
Um, no, actually AI makes it better because the cost is lower now. I'm not sure what point you're trying to make here, obviously organizations already fight against tech debt all the time through a variety of means?
76.
by pu_pe 4mo ago
That's one dimension of it, but in the context of this thread we are talking about how maintainable a codebase is for other humans. If your codebase is messy you depend on a few key employees and it might be hard to onboard new ones, s
77.
by pu_pe 4mo ago
Absolutely. I really don't think the future will be humans reading and picking apart an AI-generated codebase, there will be tech debt agents or whatever running overnight.
78.
by pu_pe 4mo ago
You are assigning intentionality to these mechanisms, but as the other commenter pointed out a much simpler way to view it is that death and senescence are just side effects, not actual mechanisms with evolutionary purposes. Organisms get b
79.
by pu_pe 4mo ago
> Please correct my math/assumptions 1) You forgot to include the actual costs of the GPUs and other equipment, you are just calculating launch costs which will always come on top of the other stuff 2) Datacenters require repairs an
80.
by pu_pe 4mo ago
Here is another scenario. You mobilize your local community or even country to choose option 1, communal ownership. Then another country or region follows path 2. If option 2 is more productive (maybe because you redirected productivity gai
81.
by pu_pe 4mo ago
This is systematic fraud, and anyone trying those antibodies with falsified data will waste money and time. A lot of papers have been retracted for similar issues. Thermo Fisher is a major worldwide supplier of antibodies, so this has quite
82.
by pu_pe 4mo ago
No marketing or design company would duplicate a band from another experiment (taking care to rotate it to make it look different nonetheless). Even in that unlikely scenario, Thermo Fisher is still responsible for the scientific data they
83.
by pu_pe 4mo ago
I agree with you in general, but depending on the task I also find that a certain level of encyclopedic knowledge can be very valuable. For example, if you use it for coding, the model will likely not resort to search or RAGs when deciding
84.
by pu_pe 5mo ago
Far larger and less populated. The point of the comment is to illustrate that the TAM is small and shrinking rather than the opposite.
85.
by pu_pe 5mo ago
The stock market doesn't operate on long-term principles anymore. So, in some sense, it is immaterial that the rosy scenarios where AI is responsible for >$30 trillion in the next decade are unlikely. If you bet against the hype and
86.
by pu_pe 5mo ago
Wouldn't the same argument apply against Tesla?
87.
by pu_pe 5mo ago
There is no substitute for effort and domain expertise, that's my point. Polymarket would like you to believe that their odds represent some sort of collective knowledge, but they don't, it's not like the world's top exp
88.
by pu_pe 5mo ago
Smaller markets are easily swayed by not so much money. Arbitrage is difficult because there is an inherent threat that you're betting against insider information. In this specific case it's clear that the market doesn't repr
89.
by pu_pe 5mo ago
Even for that function, I think their benefits are overrated. Someone could easily manipulate that number if it fits their purposes (ie a health tech startup trying to peddle some cure). The numbers given to you are an illusion of understan
90.
by pu_pe 5mo ago
The problem with measuring AI productivity is that the people doing the actual job (paralegals, developers, etc.) are doing it for someone else (judges, managers, etc.). More work, or even a speedup does not actually benefit them. So when y

More ›