Pablo Hoffman

I just met John Carmack! And he gave me forty bucks! @ID_AA_Carmack

@javisantana Has probado shortwave.com?

Mucha IA, mucho video y demás pero no tenemos un lector de email “LLM native” que me quite toda la broza que nunca pedí y nunca he leído

Looks like bots are now vulnerable to social engineering too. MitnickGPT launching any time soon?

@minimaxir Isn't this the fundamental compromise between (opinionated) frameworks and libraries though?. We had very similar tradeoffs when building Scrapy.

The Problem With LangChain

@victorianoi @graphext @HoxtonVentures Enhorabuena Victoriano!

TLDR: We are excited to share that @graphext has just closed a $4.6M seed round, lead from @HoxtonVentures and 80 angels, bringing our total seed funding to $7M 🎉 graphext.com/post/graphext-…

I often argued that, as humans interact more and more in text chat forms (whatsapp, telegram, slack) our ability to speak (the classic form of interaction) will begin to atrophy. This atrophy will likely be accelerated by the adoption of generative AIs for language.

@arankomatsuzaki @MetaAI @memdotai mem it!

REPLUG: Retrieval-Augmented Black-Box Language Models REPLUG with the tuned retriever significantly improves the performance of GPT-3 (175B) on language modeling by 6.3%, as well as the performance of Codex on five-shot MMLU by 5.1%. arxiv.org/abs/2301.12652

14 years ago, Sam Altman was the odd one on this list: paulgraham.com/5founders.html

Yendo a @PuntaTech, si alguien de mi red anda en la vuelta avisen, me encantaría (re)conectar!

🔴 CHAT-GPT y PROGRAMACIÓN Uno de los ejemplos más interesantes de ayer fue el siguiente: pedirle a GPT que nos programara el juego del TicTacToe en Python. En segundos el sistema generó un código funcional que llevado a Google Colab nos permitía echar un partida. Sigo [1/n]🧵

We're one step closer to The Entire History of You (Black Mirror episode): techcrunch.com/2022/11/01/rew…

Wow, Uruguay is the 3rd country with highest purchase power, after Switzerland and Norway (no surprise with those two), according to the Big Mac Index. ... and the first one adjusted by GDP. Go figure!

Today is the 1 year anniversary since @GitLab went public. Here’s an overview of how we prepped for our S-1 filing and setting the IPO process in motion. Thank you to everyone who helped us along the way! bit.ly/3Vrpyai

@Proxyway1 @ExtractSummit Thanks Adam, good recap!

This year's @ExtractSummit #webscraping conference was a blast.💥 If you couldn't attend it, Zyte has made the videos available on demand. Take a look at our recap to see what they're about: proxyway.com/news/zyte-2022…

@paganzacom por qué no publican la app en el Play Store fuera de Uruguay?. Ya que es una gran aliada para uruguayos que residimos en el exterior. Soy gran fan de la app desde los comienzos, no me defrauden! cc @sfogel

@Samplia máquina Parque Sur. Recabaron mis datos personales y nunca me entregó nada, si bien la máquina estaba llena. Que fraude..

My question on the best tools for web scraping is getting good engagement (on HackerNews front page for the last 3 hours): news.ycombinator.com/item?id=324096… Do chime in if you have anything to say.

So much practical advice condensed in ten tweets that I feel compelled to share it, even when it comes from the founder of a (Zyte / Scrapinghub) competitor 😉

Good collection of Scrapy practical guides in The Scrapy Playbook by @IanJKerins: thepythonscrapyplaybook.com

Elon Musk is asking about Ghislaine Maxwell's client list. He isn't shying away from dangerous topics. Interesting.