One could also phrase this principle in terms of entropy (in the information-theoretic sense): If I read some code and come across filter() and map(), I already have a pretty good idea of what the code is doing, without even looking at the filter criterion or the mapping function. Entropy is low. (There's only so many filter and map functions you could write for any given type or pair of types.)
Meanwhile, if I see reduce(), "anything" could happen. (Well, of course not anything but the set of possible reducers is surely much larger.) So entropy is high.
Avoid high-entropy constructs in your code. Try to keep entropy as low as possible. (For the same reason, code with a principled approach regarding side effects is a lot better than code where any function could mutate global state at any given time.)
Agreed! Many people are saying AI isn't really intelligent yet because it can't come up with genuinely new things. Maybe finding a rigorous formulation of QFT / high-energy physics would be a great test for whether they are!
> The “urine-no,” a hostile antiurination surface:
> Instead of minimising splash, it can instead be maximized by setting […], see Fig. 7a. Although not suitable as a commonly practical urinal, it showcases the design philosophy. Such a surface could be installed outdoors to deter public urination as the offender would fall prey to enhanced splash-back. This hostile surface may be dubbed as “urine-no,” as shown in Fig. 7b.
I spent this past weekend giving UTM w/ Apple Virtualization a shot and can confirm that it really does work fantastically well! The only downside seems to be that it's not possible to pass USB devices through to the VM guest yet?
> but the latter is eroding very fast and they're roughly the same.
Citation needed. Some politicians keep on pushing for data harvesting and secret backdoors and what not but most attempts have been blocked and we're still nowhere near the level like, e.g., in the US.
> Notably, Hy4 preview also contributed to its own development process, participating for the first time in the automated optimization of training methods, data strategies, evaluation frameworks, and low-level operators. The model proposed approaches, ran experiments, and iterated based on the results, with the resulting code, logs, and feedback feeding into subsequent rounds of exploration. This established an early-stage recursive self-improvement loop.
This reminds me of one of the predictions from https://ai-2027.com/ . Only that there it's "OpenBrain" doing this, not the Chinese. And the authors of that paper were also slightly wrong about "Mid 2026: China Wakes Up": China woke up already a while ago. And:
> But China is falling behind on AI algorithms due to their weaker models. The Chinese intelligence agencies—among the best in the world—double down on their plans to steal OpenBrain’s weights.
No need to steal anything, they have already caught up.
And then there's this prediction for February 2027:
> Officials are most interested in its cyberwarfare capabilities: Agent-2 is “only” a little worse than the best human hackers
If the distillation "attacks" created useful inputs to open weight models, ai-2027 was directionally correct that the Chinese would find ways to extract IP from western firms. (Scaled account creation and grinding outputs etc is not a dramatic story element as spies, though!)
Whether the distillation has constituted "attacks" or has or will meet the bar of "stealing" IP is not super interesting to me, though.
The idea that distillation is a significant contributor to the capabilities of the Chinese models is not true. Kimi K3 came out 2 weeks after Fable and uses a number of novel NN architecture innovations.
That may be true, though Anthropic reported 3.4 million exchanges with Moonshot months before Fable.
Why bother creating hundreds of (presumably paid) accounts and the tooling to create and consume the data if not of tangible value to their core mission?
Because the founding pillars for most of these labs are basically "more data can't hurt" and "no one died from too much data" and "you can never have enough data".
It's actually turbo boring and predictable. Capitalism has long since standardized on out and out lies to influence public perception and government action.
Don’t need a “better” hacker if you have ten thousand AIs all trying literally every single possible thing to exploit a system with. The main issue is that this will eventually bring down the exploitation cost enough to target very minor targets who weren’t worth it before.
The AI 2027 paper/website is exactly the same as random guesses from tech bros after a couple of beers telling you what they think the future will be. It has nothing to do with political theory, economic theory, game theory, or any other quasi-scientific or rigorous evaluation of real world events and predictable outcomes. It's just vibes. If they're wrong nobody will notice, if they're right people will call them geniuses.
Meanwhile, if I see reduce(), "anything" could happen. (Well, of course not anything but the set of possible reducers is surely much larger.) So entropy is high.
Avoid high-entropy constructs in your code. Try to keep entropy as low as possible. (For the same reason, code with a principled approach regarding side effects is a lot better than code where any function could mutate global state at any given time.)
reply