_heimdall

7 ms·

1.
by _heimdall 3d ago
I agree the concept isn't really important on the safety front. I feel the same way about debates whether an AI can be conscious or sentient. Those debates devolve mostly into definitional disagreements.
2.
by _heimdall 3d ago
Plenty of apes own revolvers, and yes we shoot stuff with them for fun.
3.
by _heimdall 3d ago
Oh I completely agree the tests should be entirely air gapped. If you went back 5ish years and told anyone in AI research tests with models on this scale are being some without an airgap they'd be very surprised as it was common knowle
4.
by _heimdall 4d ago
They aren't aligned, that's the problem and I don't think its a solvable one. They may have learned from humans, but they aren't aligned with us. That has all the usual questions like which humans they're aligned wi
5.
by _heimdall 4d ago
Why assume that because you haven't seen a model or an agent that none of them do? No one I've met has murdered anyone as far as I'm aware, but that doesn't mean no one has murdered another person. I also don't know
6.
by _heimdall 4d ago
Political, social, and legal options focus on a different problem, he calls that out a paragraph or two later. > Risk management is not just about cybersecurity, corporate responsibility or regulation, although those matter too. What you
7.
by _heimdall 4d ago
That does sound simple, but how can you be so sure? They never bothered to find a way of actually understanding what happens during inference. All we can do is guess, and while your explanation seems reasonable we can't actually know,
8.
by _heimdall 4d ago
> LLMs do not desire That seems likely, but we have no way of knowing this. The only real insight we get into LLM "thought" is the human readable text they produce as chain of thought. Reading it at face value it can seem to in
9.
by _heimdall 4d ago
Alignment isn't a solvable problem, the fact that guardrails are needed in the first place proves that.
10.
by _heimdall 6d ago
The Surface Duo really was begging for Windows Phone software. They never should have bothered with the Duo after killing off their mobile OS.
11.
by _heimdall 7d ago
How does this relate to Tailwind though? Tailwind does ultimately compile down to CSS. Anything you can do in Tailwind can be done in CSS.
12.
by _heimdall 7d ago
It's inline styles, not online stylesheets. Though yes, inline styles are limited.
13.
by _heimdall 7d ago
It was/is inline styles, not inline CSS. It isn't a stylesheet at all, just a single style property.
14.
by _heimdall 7d ago
Why are your buttons `a` tags?!? Buttons should be `button` tags. Links should be `a` tags.
15.
by _heimdall 7d ago
I don't see any reason the business model shouldn't exist, though yes their scale was a challenge IMO. Selling off the shelf, tested components and themes isn't a bad business. I'd firmly put it in the lifestyle business
16.
by _heimdall 13d ago
Its hard to call that the largest electric aircraft IMO. It's electric until it can't, then its not.
17.
by _heimdall 14d ago
This may be a difference in location or urban vs rural? I live in a more rural area and many people I've hired over the years own their equipment (well, if having a loan on it counts). That goes for tractors obviously, but similarly fo
18.
by _heimdall 14d ago
I don't read that recommendation as prepping. If I were to start an earthworks business I likely would rent heavy equipment at first, but once I get the business going I'd likely begin to buy my own machines. That's not prepp
19.
by _heimdall 19d ago
Oh we do agree there, control is more practical. I don't personally think alignment is even possible. The problem with control is that it will fail at scale. We can't control something that is actually smarter than us, if AI (LLMs
20.
by _heimdall 20d ago
My argument isn't whethwr factionalism is a part of us. My argument is that different faction often, or even just sometimes, aren't aligned with each other. Alignment isn't inherent in the system. Sometimes groups are aligned
21.
by _heimdall 20d ago
No, regulations help control they don't help align. I'm not sure what AI and heavy machinery have to do with each other, but I may just be missing a connection there.
22.
by _heimdall 20d ago
The parent was referring to adding `link` tags in the `<head>`, not in HTTP headers.
23.
by _heimdall 20d ago
The ideas is that the URL references the resource and the content type requested is only asking for that content in a different projection or representation. The content at a URL should always match, the format in which its represented can
24.
by _heimdall 20d ago
For the LLM use, the challenge is that it will only discover those after first requesting and parsing the HTML version. Maybe it will notice those, and maybe it will figure out the pattern for follow-up page requests, but there's no gu
25.
by _heimdall 20d ago
Its not aligned either though.
26.
by _heimdall 20d ago
People in similar cultures and societies tend to roughly align, but its still leaky within a society and can be very misaligned across different societies. The problem of alignment with AI is that there's potentially a power imbalance.
27.
by _heimdall 20d ago
Regulations are for control and punishment, not alignment.
28.
by _heimdall 21d ago
Are they? Humans are often at war with other groups of humans. And they do absolutely terrible things to the "other" group.
29.
by _heimdall 21d ago
We don't align humans. Just look at how often in documented human history there weren't wars going on somewhere.
30.
by _heimdall 21d ago
I feel for the team having to deal with these problems at github today. Its also beyond me how those in charge for years wouldn't have seen this coming. Github was bought by Microsoft, whether they wanted to acknowledge that internally

More ›