19h

9 ms·

61.
by 19h 3y ago
I’m doing reverse engineering work every now and then and a year ago I’d have called myself a fool but I have found multiple exploitable vulnerabilities simply by asking an LLM (Claude refuses less often than GPT4, GPT4 generally got better
62.
by 19h 3y ago
Pro-tip: ask Claude to "rewrite this article in a cynical tone" and you'll be never go back. Changed my entire consumption of news.
63.
by 19h 3y ago
Make those files torrents and I’ll make sure your data will transcend into the unenforceable infinities of the internet.
64.
Fractal Brownian Motion: Simulating Realistic Terrains with Fractal Noise(thebookofshaders.com)
1 points by 19h 3y ago|0 comments
65.
Onion Futures Act – US law banning trading futures on onions(en.wikipedia.org)
14 points by 19h 3y ago|2 comments
66.
by 19h 3y ago
These new models get released in such an extreme cadence now it's almost impossible to stay ahead of the curve in production deployments.. Absolutely can't wait to test drive this one -- although I'm pretty sure my 96GB M2 Ma
67.
by 19h 3y ago
Not sure how long Amex is going to be the worse option when Visa and MC are raising their fees, though
68.
by 19h 3y ago
What are those “tasks in your pipeline”? We dedicated an entire team of 6 for two months on evaluating LLMs and while Claude 2 was the final choice, we found Llama 2 70b to be absolutely great (for summaries and structured data generation f
69.
by 19h 3y ago
I’m from Germany — we drove to Arizona from California a few years ago and were too late for check in which was super annoying. I got a SEIKO Astron watch that synchronizes the time via GPS after that and couldn’t be more satisfied; it also
70.
by 19h 3y ago
I was trying to find out about video stabilization options on Linux yesterday and stumbled over an article that extremely confidently was discussing a tool that’s maintained by the „authors of ffmpeg“ … which surprised me. So I googled it a
71.
by 19h 3y ago
Talked to a few LinkedIn engineers a while ago and was astonished at the size of some payloads they’re pushing over Kafka.. in almost all production deployments I’ve done with different teams in the past once we got rid of the big payloads
72.
by 19h 3y ago
Absolutely the best font I’ve ever used. I can easily read letters from far, from close, in bright light, in darkness.. used it since day one. Also using it in my terminal. Thanks JetBrains, I love it.
73.
by 19h 3y ago
I signed up again with "country" being United States and I got it immediately -- apparently they're not auto-accepting other countries..
74.
by 19h 3y ago
Interesting idea! Can you elaborate a bit more?
75.
by 19h 3y ago
Apologies, that's a very semantic folding specific term -- with retina I refer to a word -> fingerprint database that gives you either fingerprints of single words or of blocks of text (by applying relevant heuristics to filter out
76.
by 19h 3y ago
Yes, we don't actively cover this. Our training dataset is composed of 12 different languages. The English corpus has a size of around 25TB -- ~18TB after a rather aggressive filtering process (char frequencies [i.e. https:/&#x2F
77.
by 19h 3y ago
Care to elaborate? Not sure why this tone is appropriate. The 92% is an average and not the exact accuracy across all categories; the accuracy varies by category as every category is represented by its own filter.
78.
by 19h 3y ago
We also trained on all of pushshift and have an average ”unknown” word rate of less than 0.007% — the Reddit corpus is rather amazing to capture pretty much all misspellings of a word. We may only be using 16k vector values but that doesn’t
79.
by 19h 3y ago
Unfortunately, I can't. We have some projects bubbling around that may see the light of the day eventually but given the myriads of NDAs that stack on top of each other this is rather unlikely. That said, here's some reading mater
80.
by 19h 3y ago
We’re classifying gigabytes of intel (SOCMINT / HUMINT) per second and found semantic folding or better in classification quality vs throughput than BERT / LLMs. How it works — imagine you’re having these sentences: “Acorn is a tr
81.
by 19h 3y ago
It’s improved a lot the last year. Check it out again
82.
by 19h 3y ago
Using it everyday and I can’t confirm this at all. Worst case you can enable clippy to run in the background.
83.
by 19h 3y ago
I actually got the PhD so it’s not all ways leading to failure — just accept when your advisor offers you to take time off.
84.
by 19h 3y ago
Unrelated to the actual site — we’ve been testing a tuned version of airoboros-64b on a H100 cluster as a drop in replacement for Claude-100k and GPT4-32k. It’s performing rather well in the generation of text but the 2k context definitely
85.
by 19h 3y ago
I’d be careful with H100 on clouds right now — we ran Onnx & PyTorch payloads an ran into issues that haven’t been addressed yet. For instance, onnx & PyTorch just died randomly and a few thousand dollars later we figured it was dir
86.
by 19h 3y ago
„These estimates are derived from the number of directory requests counted on directory authorities and mirrors.“ Depending on how these numbers are obtained, there is a non-zero chance at least part of this increase is caused by us.. note
87.
by 19h 3y ago
Try falcon-40b-instruct, that’s optimized for better interaction with instructions and performs A LOT better
88.
by 19h 3y ago
"You can talk with me anywhere." -- and yet no Telegram or Signal :-)
89.
by 19h 3y ago
falcon-40b also outperforms GPT3.5 -- is it powering Pi.ai?
90.
by 19h 3y ago
Something I find hilarious is that the visionOS disk image is called “xrOS” (the name that leaked before the keynote).

More ›