wavemode

20 ms·

361.
by wavemode 8mo ago
> You can do it via the C ABI, and use opaque pointers to represent higher-level Rust/C++ concepts Yeah but, you can do the same in Swift
362.
by wavemode 8mo ago
You do need to gate the web browsing. 2FA and/or credential storage helps with passwords, but it doesn't help with other private information. If the claw is currently, or was recently, working with any files on your computer or an
363.
by wavemode 8mo ago
Presumably, the same MCP tool allowing it to read specific definitions would also allow it to write out specific definitions (rather than, again, having to write out the whole file just to make one change). Otherwise it defeats the purpose.
364.
by wavemode 8mo ago
Presumably, more than one person worked on this video.
365.
by wavemode 8mo ago
The difference there would be that they would be guilty of theft, and you would likely have proof that they committed this crime and know their personal identity, so they would become a fugitive. By contrast with a claw, it's really yo
366.
by wavemode 8mo ago
The mere act of browsing the web is "write permissions". If I visit example.com/<my password>, I've now written my password into the web server logs of that site. So the only remaining question is whether I can be
367.
by wavemode 8mo ago
The LLM doesn't need to read the whole file in example 2). Why would it?
368.
by wavemode 8mo ago
I still don't follow. Why doesn't that count as cost?
369.
by wavemode 8mo ago
I mean, it depends on whether it's the inclusive "we". I'm sure it is a bigger problem than the people who worked on the video thought.
370.
by wavemode 8mo ago
(2024) past discussion: https://news.ycombinator.com/item?id=39313623
371.
by wavemode 8mo ago
I think one could easily build an MCP tool wrapping LSP which smooths over those difficulties. What the LLM needs is just a structured way to say "perform this code change" and a structured way to ask things like "what's
372.
by wavemode 8mo ago
Not sure I follow. You seem to have omitted the part of 1) explaining how the LLM knew that my_function even existed - presumably, it read the entire file to discover that, which is way more input tokens than your hypothetical available_f
373.
by wavemode 8mo ago
I hope that someday LLMs will interact with code mostly via language servers, rather than reading the code itself (which both frequently confuses the LLM, as you've noted, but is also simply a waste of tokens).
374.
by wavemode 8mo ago
> Lua is an example of such a language, and when a semicolon is necessary is when you have something that could be misconstrued as being a call: (function() print("Test1") end)(); -- That semicolon is required (functi
375.
by wavemode 8mo ago
> unique IDs to each subatomic particle You only need one ID for each type of particle. Since the laws of physics dictate that the particles themselves are indistinguishable from each other.
376.
by wavemode 8mo ago
What PRNGs lack compared to TRNGs is security (i.e. preventing someone from being able to use past values to predict future values). It's not that they somehow produce statistically invalid results (e.g. they generate 3s more often tha
377.
by wavemode 8mo ago
Yeah that's the part I'm not following. You think AI researchers would have their life's work invalidated by the creation of AGI? How? Presumably (in that scenario) their life's work (AI research) will have been foundati
378.
by wavemode 8mo ago
Arc is already the name of a programming language: http://arclanguage.org/ In fact it's the language HackerNews is written in.
379.
by wavemode 8mo ago
Consensus can be faulty yet still affect the system. If a large number of investors are expecting companies to replace workers with AI, then that's exactly what many CEOs are going to (try to) do. Since the CEO's boss is the share
380.
by wavemode 8mo ago
Did your eyes stop reading the "hard work" comment once you got to the word "and"? Preparing for chess tournaments is always hard work. But the point is that he no longer has motivation to do that hard work to partic
381.
by wavemode 8mo ago
You and I are saying the same thing.
382.
by wavemode 8mo ago
> expert opinions can be the slowest to update on the news that their specific domain may have, in hindsight, have been the wrong horse. It's the quote about it being difficult to believe something that your income requires to be fa
383.
by wavemode 8mo ago
I had a different issue with the metaphor - shouldn't OpenAI be the empire ? The death star would be the thing they created, i.e. ChatGPT.
384.
by wavemode 8mo ago
We often think of chess as something you "learn" how to do. But players like Magnus are evidence that there's really some neurological "muscle" for chess which some people just have naturally more of than others. Th
385.
by wavemode 8mo ago
"from first principles" has been a common phrase in science and philosophy for a long time: https://en.wikipedia.org/wiki/First_principle
386.
by wavemode 8mo ago
tbh that's the least surprising aspect of this. Most journalists do not have work-life balance.
387.
by wavemode 8mo ago
I don't think Git/GitHub is really all that big of a lock-in in practice for most projects. IMO Git is not an unassailable juggernaut - I think if a new SCM came along and it had a frontend like GitHub and a VSCode plugin, that al
388.
by wavemode 8mo ago
I'm envisioning that in Rust (and Python), the operator overload would be on a class/struct. It would be the macro/decorator (the same one that adds logging) which would turn the function definition into an object that implem
389.
by wavemode 8mo ago
> The Killer Feature: |> with Auto-Tracing. No other language has this combination Of the languages listed, Elixir, Python and Rust can all achieve this combination. Elixir has a pipe operator built-in, and Python and Rust have operat
390.
by wavemode 8mo ago
> I think it all boils down to, which is higher risk, using AI too much, or using AI too little? This framing is exactly how lots of people in the industry are thinking about AI right now, but I think it's wrong. The way to adopt ne

More ›