esjeon

22 ms·

241.
by esjeon 3y ago
Yeah, it doesn't make sense for the team to keep hosting every single state of the distro, just to boost building some random dudes' old/peculiar setup. It's almost like Nix team is providing an artifact hosting service.
242.
by esjeon 3y ago
C-interop is pretty perfect for most usecases. The only trouble I had was that, since command line arguments are converted to slices in Zig, I have to convert them back to null-terminated strings whenever I call C functions using any of the
243.
by esjeon 3y ago
I'm not sure if this is a joke or serious.
244.
by esjeon 3y ago
This is anyway a niche market on Linux - you can go really far with gnome/kde default terminals, and even slimmer ones like xterm/urxvt/st are more than enough to perform all the necessary tasks. Perhaps, on Mac, you badly ne
245.
by esjeon 3y ago
It is still harmful in the sense that it gives a false sense of safety. See those lethal accidents where drivers were totally distracted.
246.
by esjeon 3y ago
Yeah, but software is complex, and we don't have tools to effectively analyze its code. The scanning solutions currently available in the market are really crude, and most of them perform behavioral analysis looking for very basic vuln
247.
by esjeon 3y ago
Rust has been a practical language since the beginning. Not every experimental stuffs are academic , especially when concepts are not rigorously formulated. Rust has been a rolling language ever since it was created, so is hardly formal.
248.
by esjeon 3y ago
Caching can delay the net neutrality issue by NOT serving the contents from the first party. ISPs peering Netflix won't suffer the heavy traffic (thus no reason to charge tons of money), and ISPs that hosts caches can achieve higher Qo
249.
by esjeon 3y ago
If it gets fully open sourced, attackers can use it to find its holes more efficiently using automated tools.
250.
by esjeon 3y ago
I think the analogy doesn't work here. In my understanding, LLMs already hit a big wall. We can't increase the size of models mainly because it's too expensive, but also doing so may not be as effective as before. We've
251.
by esjeon 3y ago
> a capacity for common sense. Uh, nope. Being able to spur out text is far from understanding what common sense is. If it did have the common sense, why would OpenAI struggle so much with filtering? Because the model doesn't compr
252.
by esjeon 3y ago
I think the federation part of this post is kinda ill-written. BGS is more of a proxy for messages in PDS, and not critical for federation itself. It's something additional to the federation network, but still necessary to build a user
253.
by esjeon 3y ago
I guess some may have leftover funds from the pandemic period.
254.
by esjeon 3y ago
I don't think most normal people will spam their own SNS/blogs using LLMs. If you hate writing, you won't be writing at all, instead of using LLMs to generate stuffs. If you have a message to send, you'll be taking much
255.
by esjeon 3y ago
Oh, this is why this news felt very old and wrong. the "0.01 QPS" number for ripgrep is just a on-paper math, not an actual performance number from a running system.
256.
by esjeon 3y ago
Apple did exactly that on their keynote for M1 by comparing it to an old AMD. I really hated it because it's not quite Apple to pull cheap tricks like that.
257.
by esjeon 3y ago
Yeah, but writing a big chunk of new code always involves either gambling or cargo culting, until you nail the actual requirements and the design. MSA is just a methodology to contain risks from the uncertainty, and it never says you must b
258.
by esjeon 3y ago
I hate to admit, but Python, C, Java, and JS cover most of the modern programming. But not supporting C# sounds like a bad idea.
259.
by esjeon 3y ago
The title is obviously too hyperbolic, as the article itself doesn't even try to back the claim. It's just a good introduction piece. (The official doc is a little bit too scattered to read through.) Although I like the basic conc
260.
by esjeon 3y ago
In case of st, it actually delays the rendering. Its default config says min-latency is 8ms, and max-latency is 33ms. This probably is why OP got 20ms as an average value. The thing is that suckless people simply want to render less, and th
261.
by esjeon 3y ago
It's more like ChatGPT won't replace most Graphical -UIs. Obviously, human is much more capable at processing visual information, so even in the era of LLMs (which is yet to come) we still have to rely heavily on good GUIs. Nothi
262.
by esjeon 4y ago
> That's not where the goalposts are for the definition of AI. We've well past that goalpost, like decades ago, if you haven't realized. The simplest form of AI is just a series of if-else statements (expert system). A hig
263.
by esjeon 4y ago
This is nothing significant TBH. First, the image sucks. Do you really get any artistic whatever from that image? At best, it looks like a badly photoshopped image that struggles to simulates 80's look and feel. It also has AI-specific
264.
by esjeon 4y ago
It's a big sigh considering that the specific category is already open to Photoshop.
265.
by esjeon 4y ago
The OpenAI GPT-4 paper itself.
266.
by esjeon 4y ago
Nah, RLHF is what made GPT-4 outperform 3.5. The base model hasn't been much improved since 3.5. Also, the calibration curve is based on a subset of MMLU, so it doesn't really represent any of the actual user experience.
267.
by esjeon 4y ago
This is gold. Really. Well balanced, and much details.
268.
by esjeon 4y ago
Yup, this. I found many ex-Googlers who just don't know how to count up . In the past, "we don't know how to count that low" was a valid joke, but, these days, it seems like a lot of young engineers don't really un
269.
by esjeon 4y ago
Right. If you don’t dig deeper, it will give you only generic answers, and even fully incorporate the intention of your questions, strengthening your biases.
270.
by esjeon 4y ago
Not really, but it still helps you a lot with breaking your own bias.

More ›