Palmik

27 ms·

391.
by Palmik 3y ago
It's one of the key results from the paper. From the abstract: "In contrast to demographic reconstructions based on mtDNA, we infer a second strong bottleneck in Y-chromosome lineages dating to the last 10 ky." From the body:
392.
8k years ago, 17 women reproduced for every 1 man(ncbi.nlm.nih.gov)
45 points by Palmik 3y ago|84 comments
393.
by Palmik 3y ago
Figure 3 is interesting, it's topic distribution of 100K sampled conversations: Cluster 1: Discussing software errors and solutions Cluster 2: Inquiries about AI tools, software design, and programming Cluster 3: Geograph
394.
Lmsys-Chat-1M: A Large-Scale Real-World LLM Conversation Dataset(arxiv.org)
2 points by Palmik 3y ago|1 comments
395.
by Palmik 3y ago
Very different -- Spotify significantly increased the amount of music people consume, and the amount of artists that can get easily compensated (listing on Spotify is much easier than releasing a CD). In this case: Getty is going to be payi
396.
by Palmik 3y ago
Works for me: > Frost graces the window in winter's glow, > Ravens flock amongst drifted snow. > Each snowflake holds a secret hush, > Echoing soft in ice's gentle crush. > Mystery swathed in pale moonlight, > Ev
397.
by Palmik 3y ago
To some extent yes, for generic multi-modal chat-bots this could be a problem, but there are many apps that provide tight integration / smooth tooling for whatever problem they are helping to solve, and that might be valuable to some p
398.
by Palmik 3y ago
It could mean they use Google cloud services other than GPU/TPU, like Google Oauth, etc.
399.
by Palmik 3y ago
"Two-tower" or "dual-encoder" is an established terminology, many years old.
400.
Call for less regulatory hurdles for open-source AI in Europe(laion.ai)
1 points by Palmik 3y ago|0 comments
401.
Why don't people realize that 'high software profit margins' are fugazi?(buttondown.email)
1 points by Palmik 3y ago|1 comments
402.
by Palmik 3y ago
Bing has citations, Perplexity has citations, etc. I would not be surprised if ChatGPT moves some of the plugins to the free tier as well, to stay competitive over the longer term.
403.
by Palmik 3y ago
It's really a no brainier, especially in chat setting where context is key and the generated queries / API calls need to take it into account. I do this also in my side project, example: https://dstill.ai/agent&#x2
404.
Incorrect Quotation of Richard Stallman(gnu.org)
1 points by Palmik 3y ago|0 comments
405.
NATS: Connective Technology for Adaptive Edge and Distributed Systems(nats.io)
1 points by Palmik 3y ago|0 comments
406.
Intel's 'AI PC'(reuters.com)
5 points by Palmik 3y ago|0 comments
407.
Transformers from Scratch: Building 5 Language Models of Increasing Complexity [video](youtube.com)
5 points by Palmik 3y ago|1 comments
408.
Designing Rust Bindings for REST APIs(plume.benboeckel.net)
4 points by Palmik 3y ago|0 comments
409.
by Palmik 3y ago
Yes, but that's a separate product from Bard: https://blog.google/products/search/generative-ai-search/
410.
AI SQL Accuracy: Testing different LLMs and context strategies to generate SQL(vanna.ai)
2 points by Palmik 3y ago|0 comments
411.
OpenAI GPT-3.5-turbo-instruct released
25 points by Palmik 3y ago|21 comments
412.
by Palmik 3y ago
There are many such products if you look around (e.g. [1]). Fine-tuning is probably not going be enough to retain all the information in this case, so you definitely want to augment with retrieval. [1] https://news.ycombinator.co
413.
by Palmik 3y ago
Deno is great. I wish the maintainers were open to moving some of the great functionality from the CLI into separate libraries, but I understand how that might not be desirable from business perspective.
414.
by Palmik 3y ago
The article quality is extremely low, they are skimping even on the models.
415.
Recovering original text from OpenAI embeddings(twitter.com)
1 points by Palmik 3y ago|0 comments
416.
CMD Challenge(cmdchallenge.com)
1 points by Palmik 3y ago|0 comments
417.
by Palmik 3y ago
High quality text-to-speech costs $8.25/hr at ElevenLabs. While this is reasonable for many use cases, like book narration, it completely rules out others. UnrealSpeech is nearing the quality of ElevenLabs while being $0.36/hr. It
418.
Generative voice AI, order of magnitude cheaper(unrealspeech.com)
3 points by Palmik 3y ago|2 comments
419.
by Palmik 3y ago
Yes, and that's great, but what I meant is that it can definitely be much cheaper still, especially if you do not need a model that needs to be able to do everything, but is specialized on one or a few use cases.
420.
by Palmik 3y ago
Fine tuned gpt-3.5 is 8x the cost of regular gpt-3.5 for inference, other thread where this is discussed: https://news.ycombinator.com/item?id=37494335 This may be acceptable for some use cases, bot not for others where eve

More ›