dvrp

16 ms·

211.
Transformer Circuits: reverse-engineering transformers into graspable programs(transformer-circuits.pub)
1 points by dvrp 1y ago|0 comments
212.
by dvrp 1y ago
Because Worse Is Better!
213.
by dvrp 1y ago
That's awesome; making the world a better place one title diff at a time haha.
214.
by dvrp 1y ago
...and Wozniak.
215.
by dvrp 1y ago
Congratulations man, it makes me so happy to read this and I can relate to that dopamine rollercoaster phenomenon you are talking about.
216.
by dvrp 1y ago
Could not access your Google Doc.
217.
by dvrp 1y ago
San Francisco, CA; Full-time & Onsite; Member of technical staff. I'm the cofounder of Krea ( https://www.krea.ai/ ), a startup in San Francisco building browser-based creative tools and AI systems focused on persona
218.
by dvrp 1y ago
I built this with simple text-based memory archive too. What you said is simply adding git to the equation. I tried many approaches and, to my surprise, Markdown + Git + plain-old UNIX tooling is powerful. I've noticed that anchoring
219.
Cartridges: Store long contexts in tiny caches with LLM self-study(hazyresearch.stanford.edu)
1 points by dvrp 1y ago|0 comments
220.
by dvrp 1y ago
From their repo: tl;dr When we put lots of text (e.g. a whole code repo) into a language model's context, generation cost soars because of the KV cache's size. What if we trained a smaller KV cache for our documents offline? Usi
221.
Cartridges: Storing long contexts in tiny caches with self-study(github.com)
2 points by dvrp 1y ago|1 comments
222.
Building Cursor(newsletter.pragmaticengineer.com)
2 points by dvrp 1y ago|0 comments
223.
Seaweed APT2: a real-time, interactive, streaming video generation model(seaweed-apt.com)
1 points by dvrp 1y ago|1 comments
224.
by dvrp 1y ago
Related tweet: https://x.com/peter9863/status/1933035511389548948
225.
OmniPainter: Training-Free Stylized Text-to-Image Generation with Fast Inference(github.com)
2 points by dvrp 1y ago|0 comments
226.
Text-to-LoRA: Hypernetwork that generates task-specific LLM adapters (LoRAs)(github.com)
135 points by dvrp 1y ago|17 comments
227.
by dvrp 1y ago
Did you check the latest documents from copyright.gov? They’re interesting exactly because of what you’re saying
228.
Good Writing(paulgraham.com)
2 points by dvrp 1y ago|0 comments
229.
You can try Imagen 4 at Krea(krea.ai)
3 points by dvrp 1y ago|1 comments
230.
by dvrp 1y ago
OG poster and founder of Krea here. Shameless plug and we also rushed the API integration—there was no official partnership or anything. Also, as per our benchmarks/attempts, the API is a bit flaky still (it's in public preview),
231.
Wacom drawing tablets track the name of every application you open (2020)(robertheaton.com)
108 points by dvrp 1y ago|52 comments
232.
by dvrp 1y ago
What a coincidence! I think we both commented noting the phenomenon of abundance and the repercussions for us humans at the individual level. Especially from a fulfillment and autonomy point of view.
233.
by dvrp 1y ago
This is a byproduct of abundance. In this case, abundance of cognitive ability. We say that our food sucks. Yet, our elite athletes would crush Hercules or other God-like figures from our mythology. At the same time, we suffer from obesity.
234.
by dvrp 1y ago
a way to find people is to pursue your natural interests and let “the project” become its own thing organically
235.
Airbnb built a custom media format for transparent video playback at 60 FPS(twitter.com)
1 points by dvrp 1y ago|0 comments
236.
by dvrp 1y ago
Yes! This in essence. Specifically, it’s a file that contains a list of Entity-Attribute-Value assertions in triplets. It’s called “FACTS.md” and each line represents a fact. Such as “<OP>, PASSPORT_NUMBER, <VALUE>” Then put it
237.
by dvrp 1y ago
I feel like you can do the same using a single markdown file and an LLM (e.g. Claude Code). I do it that way and then I hooked it up with the Telegram API. I’m able to ask things like “What’s my passport number?” and it just works. Combine
238.
by dvrp 1y ago
Disclaimer: Krea cofounder here so, self-plug here. However, I thought of sharing because I think the idea of prompting ChatGPT-like models through visual cues is non-intuitive. It works quite well too. We are sadly not able to offer it for
239.
Show HN: GPT Paint – instruct ChatGPT with edit shapes, notes, and references(twitter.com)
1 points by dvrp 1y ago|1 comments
240.
by dvrp 1y ago
seems like it's not a new 671 B behemoth, just the V3 MoE engine with long-context RoPE + FA2, retargeted at theorem proving. see [1,2]: [1]: https://arxiv.org/abs/2405.14333 [2]: https://arxiv.org&#x2F

More ›