This is what we call a natural experiment

You can now connect GitHub repos to deep research in ChatGPT. 🐙 Ask a question and the deep research agent will read and search the repo’s source code and PRs, returning a detailed report with citations. Hit deep research → GitHub to get started.

"Press ENTER" is dead. Try this (Nvim 0.12, experimental 🧌): require('vim._extui').enable({}) - shows a [+n] hint (instead of "Press ENTER") if messages exceed 'cmdheight' (use "g<" to see all) - messages pager is now a real buffer+window - ":" cmdline has syntax

There was a day when memory was the leading cause of security problems But today folks we are about to step into a new future, a future where everyone can contribute to security vulnerabilities

Another lesson from the GPT-4o sycophancy problem: small changes to system prompts can result in dramatic behavior changes to AI in aggregate. Look at the prompt that created the Sycophantic Apocalypse (pink sections). Even OpenAI did not realize this was going to happen.

Introducing OpenAI o3 and o4-mini—our smartest and most capable models to date. For the first time, our reasoning models can agentically use and combine every tool within ChatGPT, including web search, Python, image analysis, file interpretation, and image generation.

I'm excited to introduce {boxview} to decipher your #rstats code without painfully scrolling and losing your way between if and else or nested loops. `boxview::boxview(fun)` and you're good to go. Compare standard print vs boxview(). Which do you prefer? https://t.co/4ByU483pOv

RIP graphic design

we trained a new model that is good at creative writing (not sure yet how/when it will get released). this is the first time i have been really struck by something written by AI; it got the vibe of metafiction so right. PROMPT: Please write a metafictional literary short story

Today, we at OpenAI launched Deep Researcher and I wanted to share a deeply personal story about how amazing this tool is and how it will change the world. Trigger warning, related to cancer....1/9

📊 Our latest meta-analysis, incorporating data from over 12,000 participants across 7 countries, provides new insights into how changes in daily movement behaviors influence obesity risks across the lifespan. 🔗Full text: https://t.co/138PuCpgvd 👇Explore more in this video.

Everything you love about generative models — now powered by real physics! Announcing the Genesis project — after a 24-month large-scale research collaboration involving over 20 research labs — a generative physics engine able to generate 4D dynamical worlds powered by a physics

7 La Trobe research projects secure $11 million in @NHMRC Ideas Grant funding 👏

Got a lot of {targets}? I'm building a new workflow where you can set an active target. Debug keybindings will then auto include that target's name so you can jump into your pipeline exactly where you want—within seconds. #rstats #neovim @wmlandau https://t.co/kJwpBrZUr4

Our latest Google Labs Experiment is here! #GenChess turns your ideas into playable art pieces using Google’s Imagen 3 model. Create and play today → https://t.co/YXWK2PKZ80

A small tip: you can use this method to let avante.nvim use the GitHub Copilot "free/cheap" claude-3.5-sonnet model, provided that you first chat with claude-3.5-sonnet in VSCode using Copilot Chat, and then enable the claude-3.5-sonnet permissions.

🐑🐑🐑🐑🐑🐑🐑🐑 https://t.co/pQVIMDjs7L

The univariate approach dominates meta-analytic structural equation modeling (#MASEM). Suzanne Jak and I showed that the standard errors and test statistics are highly biased. We also showed better methods to do it. https://t.co/che0IAE9lR

I gave a presentation yesterday on using #structuralequationmodel SEM-based #metaanalysis to fit various models, including random-effects, multiplicative error, hybrid, location-scale, nonlinear, multivariate, and more. YouTube: https://t.co/904H3cArQ4 https://t.co/3KOjyp1tF3

Configuring neovim be like.....