pu_pe

11 ms·

241.
by pu_pe 11mo ago
I don't see why banks need browser fingerprinting at all. Every credit card I've had in the past 10 years required two-factor authentication. If your theory was correct, trying to do a purchase in a new browser wouldn't work.
242.
by pu_pe 11mo ago
It needs human operators and still crashes 2x more than their competitor. And they have been calling it FSD for years now.
243.
by pu_pe 11mo ago
This looks pretty cool actually, nice job!
244.
by pu_pe 11mo ago
In places I have worked at that used Databricks, I feel they chose it for the same reasons big orgs use Microsoft: it comes out of a box and has a big company behind it. Technical benchmarks or even cost considerations would be a distant se
245.
by pu_pe 11mo ago
It was obvious that there would be no space for Yann LeCun after Alexandr Wang came in. He was probably just waiting for the best time to leave. I cannot judge his research output at Meta but he failed pretty bad at the LLM race. Since so m
246.
by pu_pe 1y ago
You're totally right. Ironically I am using gpt-oss for a project right now, I think its quality is comparable to the ones I mentioned.
247.
by pu_pe 1y ago
Four independent Chinese companies released extremely good open source models in the past few months (DeepSeek, Qwen/Alibaba, Kimi/Moonshot, GLM/Z.ai). No American or European companies are doing that, including titans like M
248.
by pu_pe 1y ago
Wouldn't this same logic apply to IBM back in its heyday? The biggest risk for Google is that it is already a behemoth and it has lost a lot of what made it so special. I also think they have a huge conflict of interest when it comes t
249.
by pu_pe 1y ago
I first had this impression as well, but I can think of several complications: 1) You would have to know the exact model used, and you would need to have access to their weights 2) System prompt and temperature, not sure how this handles th
250.
by pu_pe 1y ago
I was expecting them to communicate as well, I thought that was the whole point.
251.
by pu_pe 1y ago
Excellent work, I love experiments like these. Unsurprisingly the site is also lightning fast to load.
252.
by pu_pe 1y ago
It seems that you are offering up to 120x25s, or 3000s of Sora 2 video for $19/month. The API prices for Sora 2 are $0.10/second [1]. Are you just accepting a large loss to attract customers or is there a catch here? [1] https:&#
253.
by pu_pe 1y ago
What about accuracy? Maybe I'm missing something, but the crucial piece of information that is missing is whether the labels produced by both methods converge nicely. The fact that OP had >6000 categories using LLMs makes me wonder
254.
by pu_pe 1y ago
Do you need to overwrite weights to produce the effect I mentioned above?
255.
by pu_pe 1y ago
I'm curious about it too. I think there are two bottlenecks, one is that training a relatively large LLM can be resource-intensive (so people go for RAGs and other shortcuts), and making it finetuned to your use cases might make it dum
256.
by pu_pe 1y ago
How would you define intelligence? Surely not by the ability to make a critically acclaimed movie, right?
257.
by pu_pe 1y ago
But what is the ambition for ASML here, to become a LLM provider for European businesses? To build chips that help Mistral train their models? I just don't get the synergy.
258.
by pu_pe 1y ago
I don't see any way, shape or form in which ASML needs Mistral. If they are interested in AI-based chip design, they should either partner with a leading provider or keep their cards open for buying a startup that focuses on that speci
259.
by pu_pe 1y ago
The link you shared has data from 2021, which looks completely different than what you shared - and Sweden comes at 12th, but all the other nordic countries are buried way down in the ranking. It's frankly very hard to believe that the
260.
by pu_pe 1y ago
Yes, that's how I read it too. Seems odd that OpenAI would want to get involved in this space, feels like Deepmind has a huge headstart already.
261.
by pu_pe 1y ago
Volkswagen was caught cheating on its emission data and the CEO got fired, then prosecuted. Why shouldn't that be the case here?
262.
by pu_pe 1y ago
Does this apply to existing chats as well? I hear the sound of a million lawsuits in Europe concerning GDPR violations.
263.
by pu_pe 1y ago
This is aimed at the European Union, India and Brazil, who have all recently been mulling Big Tech regulations. It seems to be the reward for the massive support Trump has been given by the tech sector. It's likely the EU will cave but
264.
by pu_pe 1y ago
So if someone attempts to run malware when I visit their page, I am legally obliged to let them run it? Absurd and absolutely non-enforceable.
265.
by pu_pe 1y ago
> [China's] reserve margin has never dipped below 80%–100% nationwide, meaning it has consistently maintained at least twice the capacity it needs (...). That level of cushion is unthinkable in the United States, where regional grid
266.
by pu_pe 1y ago
This is a big reason why most modern Western large-scale infrastructure projects get delayed and cost overruns. People making decisions treat construction as if it was cloud computing: just pay for how much you need, when you need it. Some
267.
OpenAI's GPT-OSS models benchmarks worse than DeepSeek R1 and Qwen3 235B(xcancel.com)
8 points by pu_pe 1y ago|1 comments
268.
by pu_pe 1y ago
Python and other high-level languages may actually decrease in popularity with better LLMs. If you are not the one programming it, might as well do it in a more performant language from the start.
269.
by pu_pe 1y ago
Very sparse benchmarking results released so far. I'd bet the Chinese open source models beat them on quite a few of them.
270.
by pu_pe 1y ago
I'm not so sure Stockfish is a good example. The fact it can run on an Iphone is due to Moore's law, which follows the same pattern. And Deepmind briefly taking its throne was a very good example of the Bitter Lesson.

More ›