jampekka

3 ms·

1.
by jampekka 5d ago
SmartTube
2.
A Privacy Signal the Smart Glasses Industry Can Test(nearlens.app)
2 points by jampekka 5d ago|1 comments
3.
by jampekka 6d ago
These problems were quite widely foreseen during the legislative process, and partly had already been exhibited by the earlier ePrivacy directive. I don't know whether it matters who's to blame, but my trust in EU being very effec
4.
by jampekka 6d ago
> I am once again happy that the EU is fighting practices like these via legislation. As long as you manage to navigate all the dark patterns and not accidentally give your "informed consent".
5.
by jampekka 6d ago
CLI is the era and likely gonna be until some more LLM/AI-focused interfaces appear, and they are not gonna look like human UIs. There's a reason all models use bash and code, not a GUI shell, for their internal sandboxen. Making
6.
by jampekka 7d ago
I struggle with it too, but my rule of thumb is that "the" signals the reader they should know what specific instance is referred to, whereas "a" introduces a new instance.
7.
by jampekka 8d ago
Commoditization commoditizes commodities.
8.
by jampekka 8d ago
Or if you are afraid of scooping, just don't use public LLMs? At universities, apart from admin stuff, nobody will tell you what you must or musn't use. And I really doubt university ITs would be able to run these, at least in an
9.
by jampekka 10d ago
That China's official policy is now to prefer open models and open model development may be a part of it.
10.
by jampekka 10d ago
Kinda yes. The benchmarks become part of the validation set, which means the models get slightly overfit to them if they are used as criteria for stopping the training. But a lot less compared to using them in the training data. I'd gu
11.
by jampekka 10d ago
> It's barely coherent nonsense. Lots of irrelevant details, buzz words, wrong charts or confusing phrasing. Obviously LLM output. Plenty of that even before LLMs too though.
12.
by jampekka 10d ago
Easier said than done. Saying no is not so easy. There's a bit of a weird dynamic that if you could do something that others need help with, saying no can be considered rude, but saying (even if lying) that you don't know how to d
13.
by jampekka 13d ago
Djikstra seemed to manage? https://en.wikipedia.org/wiki/Edsger_W._Dijkstra#Use_of_tech...
14.
by jampekka 17d ago
Sad that there's the obvious regulatory capture angle encouraging motivated reasoning about AI risks and how they should be tackled. Maybe it's not the best idea that potentially civilization destroying technology is developed to
15.
by jampekka 18d ago
Thousands of some of the brightest minds have worked on this problem for over a century. The million bucks is not the big deal here.
16.
by jampekka 18d ago
> WOW This. I do dislike the AI oligarchs as much as the next person, but I do find the thread full of complaining a bit depressing still. If the result holds (and it looks it does), this may be one of the, if not the, biggest things to
17.
by jampekka 18d ago
Not in the past I guess?
18.
by jampekka 20d ago
I'm not much of an artist, but I'd guess trying to paint a painting of how quicksort works, in a way someone else can grasp it too, can need a lot more reconceptualization than writing a description about it. Doing plots or diagra
19.
by jampekka 20d ago
With the last line I tried to condense (and oversimplify) about these ideas: Not all writing is thinking: That all, even a lot, of writing, or parts of writing, is such that it will develop one's thinking much. For example most stuff I
20.
by jampekka 20d ago
This is an example of the problem. I did choose to write it in the way that although I knew it can be interpreted trivially, if the reader assumes I'm an imbecile. It was a "punchy" recap of the relatively long explanation fo
21.
by jampekka 20d ago
I did think about this while writing, and I made the compromise to accept that someone will nitpick about it to hopefully drive the point better for those willing to read it charitably. QED?
22.
by jampekka 20d ago
Sure, but doing an interpretative dance or an abstract painting can make you reconceptualize the whole concept too. But we're not really pushing those tools as much as writing. Take for example a non-native writer of the language. I&#x
23.
by jampekka 20d ago
> LLMs write better than most journalists. I used to think this, but now that I've been trying to use them to help with writing, it turns out they are not necessarily that great on the deeper level. The grammar and the surface style
24.
by jampekka 20d ago
> Writing is thinking. Thinking and deciding. There have been many times when I start out writing something substantial - could be an email, a blog post, a software design document, anything - when my own views substantially changed duri
25.
by jampekka 20d ago
Some ISPs in Finland have blocked it from resolving for copyright issues, but even if you get it to resolve by changing DNS servers, it goes into a Cloudlare CAPTCHA loop. I assume it's still about the spat with a Finnish blogger.
26.
by jampekka 20d ago
Plenty of delivery robots in Finland at least, with apparently not much problems with them getting stolen or destroyed. Starship alone seems to have running operations also in UK, Sweden, Estonia and Czhecia.
27.
by jampekka 20d ago
For us banned by archive.is: https://unwall.app/www.nytimes.com/2026/08/13/travel/solar-e...
28.
by jampekka 22d ago
Seems you have to compromise on HW openness and ethics vs paranoia.
29.
by jampekka 22d ago
Also the module/import system is still quite painful/hacky. Which is partly due to the long compile times. https://docs.julialang.org/en/v1.13-dev/manual/workflow-tips...
30.
by jampekka 25d ago
It manipulates the PRNG seed in a systematic way, keeping the same token sampling distribution. https://www.nature.com/articles/s41586-024-08025-4

More ›