srdjanr

To me it's obvious that anything related to caching has little to do with the model itself. And it's nice to know from the title that you could only be affected if you used this model.

Isn't whole CS basically pure math?

And all illustrations have the AI look (they're not AI images, but AI-generated SVG or Powerpoint or something)

You'd only get plain unstyled rendering with browser support. And at that level you don't need any complexity from Hugo or Obsidian - you only need a very simple script to convert md to plain html. No config or anything to maintain.

I think a better comparison is website builders, and people still paid web devs.

It is just a band-aid.

We noticed that some equations only make sense if mass is greater than what we'd expect from what we can see. So for now we pretend that there's this dark matter that we can't see, and with it all math so far checks out.

But we don't know if that missing mass exists at all, or if some formulas are wrong. So it's definitely a "hack", but it works for now and we don't have anything better. And we do expect to find an explanation at some point.

I don't think the idea is to give agents raw data, but for it to run many queries. Humans also don't look at terabytes of raw logs.

Paying 20$ per month to double my productivity, answer many of my health questions with good accuracy, help me with DIY projects around the house and do almost as good of a job as a psychotherapist is trickle down for me.

And you don't even have to pay - there are perfectly fine free tiers and open weight models.

A few people spend all their focus and tokens on one problem, and after some time I expect one of them to get at least closer to a solution. They can either work independently, or even better they team up and create a shared scratchpad for their agents. I don't see limits to scaling this, compared to OpenAI running however many subagents per problem.

And that multiply those few people by number of different problems.

If LLMs aren't an example of trickle down, I don't know what is

Did they force anyone to focus on reviewing those results instead of other ones? People choose to focus on them for a reason.

I think it's always better to publish preliminary results early (as long as you clearly say so, and they did), and let mathematicians decide which of those are worth focusing on.

But distributed over thousands of people, I see it happening. Maybe a bit slower, but still.

I don't think a system maximizing enjoyment would be any better. Why would anyone choose to work with sewage? Or why would most developers choose to align with rest of the team or with user needs, when programming is most enjoyable?

I might avoid it because in reality the odds of it finding "something" are high, and the odds of that "something" being benign are high.

And then it's likely the options are 1) get a non-risk-free test, or 2) have constant anxiety about it (at least in my case)

You could only realistically detect this in a lab.

You need to actively listen at multiple frequencies at the same time (gsm/lte/5g have many different bands, + wifi/bt at 2.4/5/6 ghz and their bands). And you need to remove interference from all other devices to conclude that signal is coming from ledger (so any home isn't an option).

On top of that, you need to be listening at the exact time it exfiltrates data (it might intentionally wait for some time after using device), and you need to have the compromised device (which the hardest part here - you could test 100s of devices and not find anything).

So I think this could only be caught with a large scale, recurring lab testing, far from "trivial to detect".