dvt

29 ms·

541.
by dvt 3y ago
> The value is in being able to prompt AI something like "Here's my data that comes from endpoint X. Give me a table that shows this data that is also searchable." This is a billion dollar idea even if at like a 70%+ succe
542.
by dvt 3y ago
To me, this is a cute toy, but I highly doubt the AI can generate anything actually useful. Plus, I already have a code editor open, so context-switching to the browser to do "coding" feels a bit non-ergonomic. I'd also like
543.
by dvt 3y ago
> 80% of people just want the ability to write a sentence and have it run the correct SQL on their data warehouse. Even this is IMO too ambitious (due to edge cases, and multi-shot prompting being basically required, which kind of defeat
544.
by dvt 3y ago
I think more and more people are slowly realizing that LLMs aren't products in themselves. At the end of the day, you need to do something like Midjourney or Copilot where there's some value generation. Your business can't j
545.
by dvt 3y ago
Yes there is, and I'm working on a product that does this (among other things). Is this a pain point for you? Do you mind connecting over email?
546.
by dvt 3y ago
> the security of a regular paycheck Given OP, it's clearly not as regular as once thought.
547.
by dvt 3y ago
> Management believes they can replace most of these employees with LLM-based automation I don't think this is materially true, it's just a nice narrative. (And a way to entice investors, which, remember, is the CEO's job.
548.
by dvt 3y ago
> a 7B parameter model using float32s will almost certainly outperform a 7B model using float4s Q5 quantization performs almost on par with base models. Obviously there's some loss there, but this indicates that there's still
549.
by dvt 3y ago
I'm sure I missed it here on HN since it's been 5 years ago now, but I had no idea Terry Davis died (in such a tragic way too). Used to watch his livestreams occasionally. Rest in peace.
550.
by dvt 3y ago
Not sure if this is sarcastic, but quite literally every single piece of art that Midjourney, DALL-E, etc. have been trained on was scraped without the copyright owner's consent. If anything, it's the epitome of hypocrisy.
551.
by dvt 3y ago
> That's the way that nature tends to work From the double slit experiment, to particle-wave duality, to the particle zoo of the 70s, to quantum chromodynamics, to asymptotic freedom, to more exotic theories like string theory, etc.
552.
by dvt 3y ago
I beg to differ. Transformers are purely an optimization. It’s not exactly right to call everything “compute scaling” but we are still, at the end of the day, fitting polynomials. And frankly, that’s probably not what our brains are doing.
553.
by dvt 3y ago
> I couldn't even find the entry point. Ha, same. I think I was eventually trying to hook into kernel-level functions and do it that way (I was using the Android client) but couldn't get far there either, though I think it&#x27
554.
by dvt 3y ago
> there would likely be some other way of removing the pinned cert checks Yes, but it's significantly harder than flipping a bit. There's also clever ways of countering this (e.g. checksumming the public key). Of course, even
555.
by dvt 3y ago
Very clever way of doing this (though I have a feeling you could probably enforce pinning even in sandboxed mode). I remember trying to MitM Snapchat back in college and couldn't figure it out as they were also using cert pinning.
556.
by dvt 3y ago
Ran it on a "real" OSS project of mine ( https://github.com/dvx/lofi/ ), and it was stuck at 99% loading for about 30 minutes. Then, when it finally parsed the codebase, when asked anything it always retur
557.
by dvt 3y ago
I think that the idea is that if you need (operative word) to buy one dollar for two dollars (not exactly the scenario, but imo the analogy works), then your business probably sucks. This is true more often than it is false. Basically, yo
558.
by dvt 3y ago
Piggy-backing on my other comment[1], and I know everyone's on the anti-Google bandwaggon, but OpenAI also fails miserably here: https://imgur.com/a/r0BXX9G [1] https://news.ycombinator.com/item?id
559.
by dvt 3y ago
I would guess a lot of SO comments probably say that concepts (a template/metaprogramming feature) are "dangerous" or "unsafe." It probably makes the connection with actually dangerous and unsafe things (like guns
560.
by dvt 3y ago
Starting work on a product where I'll need RAG + some language model (maybe llama) and Kalosm seems interesting. However, I'd like to package the model with the app. I don't really like the new trend of on-demand downloading
561.
Tea.xyz causes a flood of spam pull requests to open source projects(github.com)
46 points by dvt 3y ago|16 comments
562.
by dvt 3y ago
You can make an argument that the economy as a whole is not a zero-sum system, but money markets certainly are.
563.
by dvt 3y ago
For a philosophy professor, I really don't get the thesis of this post. Is the point that engineers like to build stuff and most stuff they build won't be a successful startup? This seems extremely trivial and doesn't solely
564.
by dvt 3y ago
Sorry, haven't messed around too much with optimizations. I thought it was quite fast compared to Tortoise for example (where generation speed was at a 3:1 ratio).
565.
by dvt 3y ago
> Is this just a containerized version of Bark I think so.
566.
by dvt 3y ago
Bark and Tortoise work fairly well. Bark does super fast inference[1] on my M1. [1] https://github.com/SaladTechnologies/bark
567.
by dvt 3y ago
It's absolutely contradictory, did you read the original article, specifically ¶29?
568.
by dvt 3y ago
> the global population as it's not a trajectory that be sustained indefinitely It's a bit wild to see Malthusianism alive and well in 2024. He couldn't predict the industrial revolution, and we likely can't predict w
569.
by dvt 3y ago
So your argument is "suicide is good as long as (toxic) men do it"? Come on, please try to keep HN substantially more intellectually stimulating than reddit and its ilk.
570.
by dvt 3y ago
I'll be honest here, this article reeks of bias. If you actually pore through the studies, marriage (and by extension, divorce) is an extremely complicated social construct. There are plenty of studies that show the opposite: divorce r

More ›