lukev

9 ms·

31.
by lukev 7mo ago
I'd got a step further and say that in business software, named parameters are preferable for all but the smallest functions. Using curried OR tuple arg lists requires remembering the name of an argument by its position. This saves roo
32.
by lukev 7mo ago
I agree, but absence of evidence is not evidence of absence, and we currently have a lot of developers who feel very productive right now. We are very much in need of an actual way to measure real economic impact of AI-assisted coding, ov
33.
by lukev 7mo ago
So Grokipedia?
34.
by lukev 7mo ago
"People should not be allowed to post publicly on the internet" is definitely a hot take.
35.
by lukev 7mo ago
From a strictly legal perspective the two are equivalent. The fact that there are structural injustices in the system is true, but that's not a question that any answer to "what should be legal" can fix.
36.
by lukev 7mo ago
I agree with the thrust of this article, that norms and what we perceive as good or desirable extend considerably beyond the minimum established by law. But a point that was not made strongly, which highlights this even more, is that this g
37.
by lukev 7mo ago
Bluesky is very intentionally about public posting. It's a bit weird to say people "don't care about privacy" when speaking of a platform designed to amplify and distribute posts as widely and effectively as possible.
38.
by lukev 7mo ago
You can care a lot about plausible deniability and the ability to delete your own data, but it seems a bit weird to denounce a whole ecosystem as "generally a bad idea" on those grounds, when that is a deliberate anti-goal of the
39.
by lukev 7mo ago
If we're not going to ban prediction markets, we could at least make it a requirement that all bets are public and associated with a real identity. That'd go a long way towards curbing the corruption of these things, while preserv
40.
by lukev 7mo ago
Also there are a ton of other ways to skin that cat… you could vibe code a Slack plugin to make this work in like 15 minutes.
41.
by lukev 7mo ago
And most the time a bit flips it means that there's a wonky pixel somewhere in a photo, texture or video that you'd never even notice. A bit flip actually needs to be pretty "lucky" to result in a crash.
42.
by lukev 7mo ago
I had success on a similar problem by allocating native buffers for the matrices, then using a basic CUDA call. The actual work was 100x faster than my CPU baseline. The bottleneck of course was fetching & loading relevant data to memor
43.
by lukev 7mo ago
The legal intricacies of this are of course interesting and relevant. But what are the social ramifications if this kind of thing is deemed acceptable? It feels like it would effectively be the end of OSS licensing, because it's pretty
44.
by lukev 7mo ago
Clojure is awesome for LLMs (if you shim in an automatic paren balancer). But that's because it's tight, token efficient, and above all local . Pure functions don't require much context to reason about effectively. However,
45.
by lukev 8mo ago
This is a valid point. However, the conclusion depends on the fact that the people in power have the information and wisdom to make the choices required stay in power. Which has been mostly true (so far) in US history. Historically this has
46.
by lukev 8mo ago
This is a highly circular method of evaluation. It correlates "fluency behaviors" with longer conversations and more back and forth. What it notably does not correlate any of these these behaviors with is external value or utili
47.
by lukev 8mo ago
Not particularly so. But even if it were, would that justify the social cost of this kind of gambling?
48.
by lukev 8mo ago
That only makes sense in a hermetically sealed system, which this is very much not.
49.
by lukev 8mo ago
Equating being "consistently right" with having a sufficiently large stash of capital is ludicrous. "right" people will wisely take most their winnings out of a high-variance market. "wrong" people with deep po
50.
by lukev 8mo ago
Set aside the question of how we might implement this (which I grant is complex and path-dependent)... but imagine if 5% of the wealth of every US billionaire were instead allocated to research and development. Ultimately I don't think
51.
by lukev 8mo ago
Frankly I'm tired of metaphor-based attempts to explain LLMs. Stochastic Parrots. Interns. Junior Devs. Thought partners. Bicycles for the mind. Spicy autocomplete. A blurry jpeg of the web. Calculators but for words. Copilot. The term
52.
by lukev 8mo ago
Right but those things exist to prevent bots. Which this is. So at this point we're talking about participating in the (very old) arms race between scrapers & content providers. If enough people want agents, then services should
53.
by lukev 8mo ago
This is being downvoted but it shouldn't be. If the ultimate goal is having a LLM control a computer, round-tripping through a UX designed for bipedal bags of meat with weird jelly-filled optical sensors is wildly inefficient. Just sta
54.
by lukev 8mo ago
This clarifies an important point for me. The derivative of a LLM agent's capabilities (on its own) is negative. It's not that they can't do useful work -- it means that (for now) they require some level of input or steering.
55.
by lukev 8mo ago
I think the author is far too generous in thinking through the possible motives of tech leaders in the "what if they believe it?" branch. Far from embracing UBI (or any other legal/social strategy to mitigate mass unemploymen
56.
by lukev 8mo ago
I mean, I find that LLMs are quite good with Lisp (Clojure) and I really like the abstraction levels that it provides. Pure functions and immutable data mean great boundary points and strong guarantees to reason about my programs, even if a
57.
by lukev 8mo ago
"You are responsible for what you ship" is actually a pretty universally agreed-upon principle...
58.
by lukev 8mo ago
I am not sure how to define consciousness, but I can't imagine a definition that doesn't involve state or continuity across time.
59.
by lukev 8mo ago
So obviously there is a lot of data in the parameters. But by stateless, I mean that a forward pass is a pure function over the context window. The only information shared between each forward pass is the context itself as it is built. I ce
60.
by lukev 8mo ago
Yes, but the state is just the prompt and the text already emitted. You could assert that text can encode a state of consciousness, but that's an incredibly bold claim with a lot of implications.

More ›