
Wordle xxx 1/6 🟩🟩🟩🟩🟩 Get Wordle right on your first guess using the daily ⬛🟨🟩 tweet distribution kaggle.com/benhamner/wordle-…

Sumble co-founder and CTO. Learning high-quality, structured data about the world. Formerly @kaggle

Wordle xxx 1/6 🟩🟩🟩🟩🟩 Get Wordle right on your first guess using the daily ⬛🟨🟩 tweet distribution kaggle.com/benhamner/wordle-…


Right now, AGI looks more feasible than more housing in San Francisco

if they cared about student outcomes would’ve spent the $557m on gpus
new uc berkeley data science building

The “AI escape the box” experiment was comically naive in hindsight. There is no box. Instead, there’ll be embedded access to every critical system in the world

What will engineering productivity look like in a year once a high fraction of codebases are written by Claude Code/Codex and no human has a detailed understanding of these parts anymore?

if country==“AU”: await asyncio.sleep(0.52)
Interviewer: Your page loads in 80 ms in Australia but 600 ms in India. Same backend. Same code. What would you use to fix this?

Excited to share more publicly what @benhamner and I have been working on: techcrunch.com/2025/10/22/su… We are building a knowledge graph for GTM data & have early customers such as @elastic, @figma and @Snowflake

english in 2030 : python :: python in 2020 : assembly

WOW - cold inbound hiring processes are the worst I've ever seen it: - Huge volume of applications (probably triggered automatically) - Large volume of fake resumes - Extensive use of AI assistants in phone screens

You could release a crazy stat on the number of times (and number of employees) ClosedAI companies have downloaded open models from HuggingFace
All closed-source frontier labs use tons of open-source all over the stack, starting from python, @PyTorch, @huggingface all the way down to RoPE, GQA, flash-attention, or any tiny improvements released by open-source players. The whole transformers architecture (the T in gpT) comes from an open research paper and open-source model from 2017: huggingface.co/papers/1706.0… This is ok but would be much nicer if they would acknowledge the contributions & contribute back more!

Assumptions: 8 billion people * 20k spoken/written words per day * 1.3 tokens per word

This is around 5x the ~200 trillion natural language tokens humans generate every month
Google is processing 980 trillion+ monthly tokens across our products and APIs (up from 480T in May) 🤯 No slowdown in sight, intelligence is everywhere.

Kaggle launched an LLM eval product kaggle.com/benchmarks This has the potential solve the biggest challenge in the LLM ecosystem: strong and diverse evals

Getting a bunch of 429 rate limit errors with a quota exceeded message from @OpenAI's API, in spite of being well under quota and published rate limits

NLC: natural language cron
We're excited to introduce Tasks! For the first time, ChatGPT can manage tasks asynchronously on your behalf—whether it's a one-time request or an ongoing routine. Here are my favorite use cases: 1/ ChatGPT checks stock price every morning!