waldrews

17 ms·

181.
by waldrews 3y ago
These days, using such plausible sounding sarcasm is dangerous, because the LLM's will interpret it as literal knowledge (especially the online LLM's, seeing the text on a high-trust site).
182.
by waldrews 3y ago
All this has happened before and will happen again.
183.
by waldrews 3y ago
I still have a hard time understanding where the AGPL copyleft feature kicks in on a product like Citus. Are you safe just deploying it as part of your backend? Like, do you have to actually touch Citus's own code to trigger it? Wh
184.
by waldrews 3y ago
What would make it easier for me to bet on this is if a few pages made with this could be embedded in an existing Django app, in particular using Django auth. I assume many other people are in that position, with existing Django assets. I
185.
by waldrews 3y ago
Software as we know it, at least enterprise CRUD and the like, will become cheap to make with fewer programmers. The AI's are going to help millions of organization who don't know what they're doing create lots of complex, b
186.
by waldrews 3y ago
Nicely done, but I suspect the 'wall of code' the kids see when first opening the page is intimidating. Maybe start with a smaller example?
187.
by waldrews 3y ago
Respectfully, I'd like to nominate "Heck, even if you want it more easy, you could write a tool in less than 100 lines that adds packages names to .json/.yaml/.toml, which allows you to parse it using the nix language an
188.
by waldrews 3y ago
So this is a Javascript engine that can be compile into WASM, so then in principle you could run it on another Javascript runtime? Whoa.
189.
by waldrews 3y ago
Previous discussion: https://news.ycombinator.com/item?id=791821
190.
by waldrews 3y ago
Curious how AGPL would apply for something like this. This seems like a tool to put a nice front end on a complex app, but would that trigger copyleft for the the overall backend?
191.
by waldrews 3y ago
Gotta love how he's like, all patent medicines are fake, except the one my friend is selling... Bah.
192.
by waldrews 3y ago
What other services fall into the "just take my money" category? I fully agree, Hetzner is a good deal.
193.
by waldrews 3y ago
Ooh and you're probably doing a split into voice and non-voice tracks of the original, and keeping non-voice at original volume, but lowering the voice track.
194.
by waldrews 3y ago
Impressively done! It sounds like you're doing 1) doing voice recognition with voice time clues, which Whisper and the like provide, breaking it up into sentence (or similar) units; you don't need to time match individual words,
195.
by waldrews 3y ago
Maybe try the Docker installation to keep the dependencies off your system: https://superset.apache.org/docs/installation/installing-sup...
196.
by waldrews 3y ago
Yes, elephants have large footprints.
197.
by waldrews 3y ago
Very cool. Looks even closer to Reflex (maybe Flet in web mode?) than Streamlit... Could someone with Reflex experience comment on the difference? Separately, all of the above seem to lack official auth solutions. People with experience
198.
by waldrews 3y ago
Elephants can swim really well, but the G-forces at liftoff would be tough on them. Plus you've gotta supply them; 70,000 Calories per day (ideally in the form of oranges), and something like 25 kg of oxygen.
199.
by waldrews 3y ago
Statistician here. There's a deep idea called the likelihood principle https://en.wikipedia.org/wiki/Likelihood_principle that says all the information we can get from the data about model parameters is containe
200.
by waldrews 3y ago
Can't seem to get it to see my NVIDIA Quadro P3000 GPU. Is it too old to make a difference?
201.
by waldrews 3y ago
So the main novelty here is billing by the second?
202.
by waldrews 3y ago
Unit testing isn't an overloaded term. Evaluation by itself is overloaded, though "LLM evaluation" disambiguates it. I first parsed the title as 'faster inference' rather than 'faster evaluation' even be
203.
by waldrews 3y ago
did you mean 'output the final result in' rather than 'input...' ?
204.
by waldrews 3y ago
But there is no unknown distribution. There is a latent parameter (the n that's uniformly sampled) whose distribution is known; so the distribution of colors for the next draw is well defined. "Depend fully on the unknown distri
205.
by waldrews 3y ago
Well, if Hacker News is a dynamical system, and discussion of the Poincaré Recurrence Theorem is one of its states, it was bound to come up again eventually.
206.
by waldrews 3y ago
Any west coast representation from Stanford/Berkeley/Caltech?
207.
by waldrews 3y ago
I was hoping it would be live and time-lapse streams of birds making nests, like https://www.youtube.com/watch?v=vtLIuEg_lDg
208.
by waldrews 3y ago
Great idea! Recommending tweaking the prompts so it uses more concise language without filler words (it seems to like starting with'The article is about...' or 'This is a discussion of').
209.
by waldrews 3y ago
Heh... I tried to write that in a 'viewpoint neutral' way, but I'd be the consultant, I'd be making them a predictive modeling piece which would help them get to MVP and get funded. So I guess I'm trying to figure
210.
by waldrews 3y ago
I hear you in the general case. The specific situation here is the startup will be much more fundable once the piece the consultant can build is done, but the consultant would rather be paid in e.g. a year than hold on to equity. So, it&#

More ›