Skip to main content

Posts

Showing posts from September, 2024

Las Fiestas Blancas de Sean Combs eran celebraciones de la alta sociedad. ¿Fueron algo más?

By Jesse McKinley and Sarah Esther Maslin from NYT En español https://ift.tt/KWnR6Dd

The Rising Risk of a New Forever War

By Fawaz A. Gerges from NYT Opinion https://ift.tt/sLFp2mi

In North Carolina, a Race to Provide Relief Amid an ‘Unprecedented Tragedy’

By Tim Arango from NYT U.S. https://ift.tt/gqW4Il1

Hassan Nasrallah, líder de Hizbulá, es asesinado a los 64 años

By Neil MacFarquhar and Ben Hubbard from NYT En español https://ift.tt/pFPa2wy

Meta Unveils New Virtual Reality Headset as It Pursues the Metaverse

By Mike Isaac from NYT Technology https://ift.tt/Wq7fLpI
By Unknown Author from NYT Magazine https://ift.tt/95kSGMn
By Unknown Author from NYT Opinion https://ift.tt/cZXJxN4

New top story on Hacker News: Launch HN: Panora (YC S24) – Data Integration API for LLMs

Launch HN: Panora (YC S24) – Data Integration API for LLMs 8 by nael_ob | 0 comments on Hacker News. Hey HN! We're Nael and Rachid, and we're building Panora ( https://ift.tt/cESKoMa ), an open-source API that connects various data sources to LLMs, from 3rd party integrations to embeddings and chunking generation. Here's a demo: https://www.youtube.com/watch?v=45QaN8mzAfg , and you can check our docs here: https://ift.tt/OJ4Qcmy Our GitHub repo is at https://ift.tt/cESKoMa . Building integrations by hand is tedious and time-consuming. You must adapt to API documentation quirks, manage request retries, OAuth/API key authorization, refresh tokens, rate limits, and data sync freshness. Moreover, you have to keep up with the constant rise of embedding models and chunking capabilities. On the other hand, with the rise of AI-powered apps, you have to handle embedding and chunking of all the unstructured data. The dominant player in this space is Merge.dev, but it has several d...

Liz Cheney Suggests a New Political Party May Be Needed

By Peter Baker from NYT U.S. https://ift.tt/fTkm5wU

What Is Hezbollah, the Militant Group Based in Lebanon?

By Ephrat Livni from NYT World https://ift.tt/lykrpoB

Pagers Belonging to Hezbollah Militants Explode Across Lebanon

By Axel Boada from NYT World https://ift.tt/mleyIoj

Chad McQueen, actor de ‘Karate Kid’, muere a los 63 años

By Alexandra E. Petri from NYT En español https://ift.tt/3DPBs6e

Donald Trump Prepares to Unveil a Cryptocurrency Business

By David Yaffe-Bellany from NYT Technology https://ift.tt/bzhD7lw

Encuestas electorales 2024: Minnesota

By Unknown Author from NYT En español https://ift.tt/sf96YmA

New top story on Hacker News: Show HN: Tune LLaMa3.1 on Google Cloud TPUs

Show HN: Tune LLaMa3.1 on Google Cloud TPUs 3 by felarof | 0 comments on Hacker News. Hey HN, we wanted to share our repo where we fine-tuned Llama 3.1 on Google TPUs. We’re building AI infra to fine-tune and serve LLMs on non-NVIDIA GPUs (TPUs, Trainium, AMD GPUs). The problem: Right now, 90% of LLM workloads run on NVIDIA GPUs, but there are equally powerful and more cost-effective alternatives out there. For example, training and serving Llama 3.1 on Google TPUs is about 30% cheaper than NVIDIA GPUs. But developer tooling for non-NVIDIA chipsets is lacking. We felt this pain ourselves. We initially tried using PyTorch XLA to train Llama 3.1 on TPUs, but it was rough: xla integration with pytorch is clunky, missing libraries (bitsandbytes didn't work), and cryptic HuggingFace errors. We then took a different route and translated Llama 3.1 from PyTorch to JAX. Now, it’s running smoothly on TPUs! We still have challenges ahead, there is no good LoRA library in JAX, but this feel...

New top story on Hacker News: Ask HN: How do you manage your prompts in ChatGPT?

Ask HN: How do you manage your prompts in ChatGPT? 10 by nabi_nafio | 6 comments on Hacker News. I use ChatGPT regularly for a lot of different tasks. For example, coding, health Q&A, and summarizing docs. The different prompts stack up in the sidebar which becomes very difficult to manage. For example, I frequently have to refer back to a prompt that I wrote previously. But I usually give up looking for it because of the tedious scroll and search process. I was wondering if there is an easier way. How do you manage your prompts in ChatGPT?

Family of American Slain in the West Bank Demands an Independent Probe

By Liam Stack and Ephrat Livni from NYT World https://ift.tt/3lGszfQ

Sabrina Carpenter Beats Travis Scott to No. 1 by a Hair

By Ben Sisario from NYT Arts https://ift.tt/S4YHOZ2

En Francia, 51 hombres van a juicio por un caso de violación con drogas

By Catherine Porter and Ségolène Le Stradic from NYT En español https://ift.tt/QDGjVkB