In some dimensions, current day LLMs are already superintelligent. They are extremely good knowledge retrieval engines that can far outperform traditional search engines, once you learn how properly to use them. No, they are not AGIs, because they’re not sentient or self-motivated, but I’m not sure those are desirable or useful dimensions of intellect to work towards anyway.

The kneejerk reaction is gonna be “Meta bad”, but it’s actually a bit more complicated.

Whatever faults Meta has in other areas, it’s been mostly a good player in the AI space. They’re one of the major reasons we have strong open-weight AI models today. Mistral, another maker of open AI models and Europe’s only significant player in AI, has also rejected this code of conduct. By contrast, OpenAI a.k.a. ClosedAI has committed to signing it, probably because they are the incumbents and they think the increased compliance costs will help kill off competitors.

Personally, I think the EU AI regulation efforts are a big missed opportunity. They should have been used to force a greater level of openness and interoperability in the industry. With the current framing, they’re likely to end up entrenching big proprietary AI companies like OpenAI, without doing much to make them accountable at all, while also burying upstarts and open source projects under unsustainable compliance requirements.

The EU AI Act is the thing that imposes the big fines, and it’s pretty big and complicated, so companies have complained that it’s hard to know how to comply. So this voluntary code of conduct was released as a sample procedure for compliance, i.e. “if you do things this way, you (probably) won’t get in trouble with regulators”.

It’s also worth noting that not all the complaints are unreasonable. For example, the code of conduct says that model makers are supposed to take measures to impose restrictions on end-users to prevent copyright infringement, but such usage restrictions are very problematic for open source projects (in some cases, usage restrictions can even disqualify a piece of software as FOSS).

At this point, Tesla needs to steal IP from them, not the other way round.

Fermentation-made milk substitute was available at supermarkets in Singapore (under the brand name Very Dairy, though the original product was from the startup Perfect Day). I really liked it—a lot nicer than oat milk for drinking straight up. Unfortunately it went off the shelves after a while, seems like demand wasn’t high :-(

I’m curious whether Deepseek will gaf about this. They’ve been rather uninterested in commercialization, and the app is mainly a way of showing off their model, which itself is released open-weights. In fact, it’s literally impossible to spend money in the app! They sell tokens but it’s API-only, and you can’t spend it in the app.

So it’s entirely possible the Deepseek will shrug, let their app be banned in Germany, and keep doing what they’re doing.

It’s a bit hard to believe, but the vast majority of China’s manufacturing is consumed in China. They’re actually not that export oriented compared to other countries like Germany or Japan, it’s just the scale that makes them such an export juggernaut. The flip side of this is that most of the energy use is also actually China’s own energy use.

And China’s energy use is increasing simply because its people are getting richer and consuming more. Based on this, I don’t think China is the main concern. There are lots more developing countries that will likewise use more energy as they develop. China’s green transition seems to be going full tilt, but I’m not sure those other countries can transition as quickly.

By that metric, you can argue Kasparov isn’t thinking during chess, either. A lot of human chess “thinking” is recalling memorized openings, evaluating positions many moves deep, and other tasks that map to what a chess engine does. Of course Kasparov is thinking, but then you have to conclude that the AI is thinking too. Thinking isn’t a magic process, nor is it tightly coupled to human-like brain processes as we like to think.

Pretty sure it’s at least semi-autonomous. In the video you can see the bots react to hits and recover their footing, there’s no way a human can control all those reflex actions in real time.

Does it? They’re a middle-upper income country now, and child labor tends to be an issue at much lower levels of development. Anyway, for the Chinese electronics sector, you’re vastly more likely to see humanoid robots than children.

That article is overblown. People need to configure their websites to be more robust against traffic spikes, news at 11.

Disrespecting robots.txt is bad netiquette, but honestly this sort of gentleman’s agreement is always prone to cheating. At the end of the day, when you put something on the net for people to access, you have to assume anyone (or anything) can try to access it.

It’s seldom the same companies, though; there are two camps fighting each other, like Gozilla vs Mothra.

It’s possible to run the big Deepseek model locally for around $15k, not $100k. People have done it with 2x M4 Ultras, or the equivalent.

Though I don’t think it’s a good use of money personally, because the requirements are dropping all the time. We’re starting to see some very promising small models that use a fraction of those resources.

So long as there are big players releasing open weights models, which is true for the foreseeable future, I don’t think this is a big problem. Once those weights are released, they’re free forever, and anyone can fine-tune based on them, or use them to bootstrap new models by distillation or synthetic RL data generation.

Power usage probably won’t be a major issue; the main take-home message of the Deepseek brouhaha is that training and inference can be much more efficiently than we had thought (our estimates had been based on well-funded Western companies that didn’t have to bother with optimization).

AI spam is an annoyance, but it’s not really AI-specific but the continuation of a trend; the Internet was already drowning in human-created slop before LLMs came along. At some point, we will probably all have to rely on AI tools to filter it out. This isn’t something that can be unwound, any more than you can undo computers being able to play chess well.

They released the major components of their training and interference infrastructure code a couple weeks ago.

Deepseek actually released a bunch of their infrastructure code, including the infamous tricks for making training and interference more efficient, a couple of weeks ago.

The strangest twist to this is that Deepseek itself seems to be the only company not trying to cash in on the Deepseek frenzy:

Liang [Deepseek’s founder] has shown little intention to capitalise on DeepSeek’s sudden fame to further commercialise its technology in the near term. The company is instead focusing the majority of its resources on model development…

These people added the independently wealthy founder has also declined to entertain interest from China’s tech giants as well as venture and state-backed funds to invest in the group for the time being. Many have found it difficult to even arrange a meeting with the secluded founder.

“We pulled top-level government connections and only got to sit down with someone from their finance department, who said ‘sorry we are not raising’,” said one investor at a multibillion-dollar Chinese tech fund.

Funny thing is, the price of lidar is dropping like a stone; they are projected to be sub-$200 per unit soon. The technical consensus seems to be settling in on 2 or 3 lidars per car plus optical sensors, and Chinese EV brands are starting to provide self driving in baseline models, with lidars as part of the standard package.

It’s strongly dependent on how you use it. Personally, I started out as a skeptic but by now I’m quite won over by LLM-aided search. For example, I was recently looking for an academic that had published some result I could describe in rough terms, but whose name and affiliation I was drawing a blank on. Several regular web searches yielded nothing, but Deepseek’s web search gave the result first try.

(Though, Google’s own AI search is strangely bad compared to others, so I don’t use that.)

The flip side is that for a lot of routine info that I previously used Google to find, like getting a quick and basic recipe for apple pie crust, the normal search results are now enshittified by ad-optimized slop. So in many cases I find it better to use a non-web-search LLM instead. If it matters, I always have the option of verifying the LLM’s output with a manual search.