Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
wseqyrku
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
15 ms
·
301.
▲
by
wseqyrku
2y ago
> (trying to move the critique beyond the title...) This is kind of a theme in HN now. The top comments are completely besides the point of the article/story/etc.
302.
▲
Algorithmic Language Models with Neurally Compiled Libraries
(arxiv.org)
1 points
by
wseqyrku
2y ago
|
0 comments
303.
▲
A Novel Neural Network Architecture for Language Understanding (2017)
(research.google)
1 points
by
wseqyrku
2y ago
|
0 comments
304.
▲
How LLMs Know When to Stop Generating?
(towardsai.net)
2 points
by
wseqyrku
2y ago
|
0 comments
305.
▲
by
wseqyrku
2y ago
Of course this doesn't need to be used on "AI use" as they frame it. So far, your activity was a line in the logs somewhere, now someone is actually looking at you with three eyes, at all times.
306.
▲
by
wseqyrku
2y ago
Yeah, this is basically a kind of surveillance system for governments seeking "insights" into communications of any modality.
307.
▲
Ask HN: How to pitch an idea without the risk of idea theft
26 points
by
wseqyrku
2y ago
|
9 comments
308.
▲
by
wseqyrku
2y ago
Yeah, the camera needs a physical lid.
309.
▲
by
wseqyrku
2y ago
For a second I thought this was buzzfeed for some reason.
310.
▲
by
wseqyrku
2y ago
> If it works on Chrome, no one cares or even tests for other things. We've been through an extensive standardization pass for this to not happen. Anything not matching the specification whether in Chrome or any other browser should
311.
▲
Better Binary Quantization (BBQ) in Lucene and Elasticsearch
(elastic.co)
2 points
by
wseqyrku
2y ago
|
0 comments
312.
▲
Can training on simultaneously generated synthetic data define consciousness?
1 points
by
wseqyrku
2y ago
|
1 comments
313.
▲
by
wseqyrku
2y ago
> dropping the part of your brand name that actually makes it recognizable This feels a lot like Twitter -> X, I guess we have to get used to this.
314.
▲
by
wseqyrku
2y ago
> Do hard things carefully Seems like my habit is making things hard and do them carefully.
315.
▲
by
wseqyrku
2y ago
> should collect funds from all UN member countries and pool them together Centralization like this just makes the matter worse in a larger scale.
316.
▲
by
wseqyrku
2y ago
Seriously, don't wrap with line breaks, whether it's a git commit message or a freaking blog post.
317.
▲
MergeLlama-7B – A fine tune of CodeLlama for resolving merge conflicts
(old.reddit.com)
3 points
by
wseqyrku
2y ago
|
0 comments
318.
▲
by
wseqyrku
2y ago
Looking at bolt.new I think all the Studio/IDE type of apps are going to look like that. Could be video or code or docs etc. I can see myself paying a little too much to have a local setup for this.
319.
▲
by
wseqyrku
2y ago
Only if you could commit the changes in the browser and pull locally?
320.
▲
by
wseqyrku
2y ago
this would be incredible for scripting.
321.
▲
by
wseqyrku
2y ago
Yeah but you have to acknowledge the distance between local compute and a gazillion dollar data center, and even past that the hardware for a local setup is premium tier so it's not gonna quite catch up considering the cost.
322.
▲
by
wseqyrku
2y ago
When you're a politician and have a business hobby
323.
▲
by
wseqyrku
2y ago
That's not what I'm suggesting. I mean this as an automated process.
324.
▲
Ask HN: Why don't we generate all the missing pages on Wikipedia?
3 points
by
wseqyrku
2y ago
|
6 comments
325.
▲
by
wseqyrku
2y ago
It's great for refactorings and text manipulation -- things I used to do regex for can be done using gpt, though that depends on the scope so it is still not quite there to replace regex. Oh did I say it's great at bash too?
326.
▲
Ask HN: What are the questions that companies run long inference sessions for?
1 points
by
wseqyrku
2y ago
|
1 comments
327.
▲
Ask HN: Can LLMs understand the passage of time if continuously running?
3 points
by
wseqyrku
2y ago
|
2 comments
328.
▲
Generative AI for YouTube
(deepmind.google)
3 points
by
wseqyrku
2y ago
|
0 comments
329.
▲
Skeleton-of-Thought: Prompting LLMs for Efficient Parallel Generation
(arxiv.org)
1 points
by
wseqyrku
2y ago
|
0 comments
330.
▲
by
wseqyrku
2y ago
Fascinating that it's just now that someone got curious about this enough to write a post. I know if you looked for it you'd probably find this stuff but it is definitely scarce IMO. Thing is, not enough developers and designers t
More ›