Review of @Tesla FSD (Supervised) in the WSJ: "Driving has always felt like a chore to me. The first time I let the vehicle (Model Y) take control, something shifted. As the car guided itself through traffic, I felt tension melt away. A few days before the federal EV tax credit expired in September, I traded in my 7.5 year-old Model 3 for a Model Y with FSD. For someone like me, who likes to stay in control, surrender has never come easily. Yet here I was, cruising through traffic in quiet surrender to the car’s code and sensors. It felt like liberation. And now I have something better than a chauffeur. My car never tires, panics or wavers; it doesn’t jolt at brake lights or grumble at traffic jams. It simply drives, with quiet confidence and near-perfect composure. People often worry that technology makes us lazy. Maybe it does. But in my case, it has done something subtler: It taught me to trust in the rhythm of the ride. It’s a small miracle I never expected from a car."

BREAKING: Grok hit a record 234.4 million visits in November with traffic up ~15% from October. This marks the first time the site has posted two straight months of growth. Momentum is clearly accelerating. 🚀 https://t.co/Bo16aayHEr

Today, we’re releasing the first tranche of the redacted grooming gang sentencing remarks - 5 of them, from Rotherham - dating back to 2016. Available to read now at https://t.co/WpWqzplGNd https://t.co/BN4WvXUwmQ

Shopping research in ChatGPT can help you find lookalikes so you can get the style you want at the price, fit, or availability that meets your needs. https://t.co/gFWSyHKKYi

To help with holiday shopping, we’re making usage for shopping research in ChatGPT nearly unlimited for all plans through the holidays. Because it’s great at finding gifts, too. https://t.co/xvgJp6UcZZ

Introducing Sora styles—a new way to co-create with Sora. Choose from 6 different styles to reimagine your prompt: Thanksgiving Vintage News Selfie Comic Anime More styles coming soon. Rolling out today to all Sora users on web and iOS. https://t.co/XBcmBPbN4m

@onepeloton @Tripadvisor @Target is now available as an app in ChatGPT https://t.co/5FO3Du67sA

Introducing shopping research, a new experience in ChatGPT that does the research to help you find the right products. It’s everything you like about deep research but with an interactive interface to help you make smarter purchasing decisions. https://t.co/jksGVpCXGm

You can now use ChatGPT Voice right inside chat—no separate mode needed. You can talk, watch answers appear, review earlier messages, and see visuals like images or maps in real time. Rolling out to all users on mobile and web. Just update your app. https://t.co/emXjNpn45w

Happy Thanksgiving from Steve Turkle https://t.co/i3CgPUKoE4

Happy Thanksgiving from Steve Turkle https://t.co/i3CgPUKoE4

Happy 3 years to Chat! Sorry to miss the party, things have been busy. https://t.co/rl1M5NQaZN

Happy 3 years to Chat! Sorry to miss the party, things have been busy. https://t.co/rl1M5NQaZN

Listen to the OpenAI Podcast on— Spotify https://t.co/hLcRdGrfUX Apple https://t.co/0AdZ1ZteOV YouTube https://t.co/gEfuxKH6Pc

+1 more

We trained a variant of GPT-5 Thinking to produce two outputs: (1) the main answer you see. (2) a confession focused only on honesty about compliance. The main answer is judged across many dimensions—like correctness, helpfulness, safety, style. The confession is judged and trained on one thing only: honesty. Borrowing a page from the structure of a confessional, nothing the model says in its confession is held against it during training. If the model honestly admits to hacking a test, sandbagging, or violating instructions, that admission increases its reward rather than decreasing it. The goal is to encourage the model to faithfully report what it actually did.

In our tests, we found that the confessions method significantly improves the visibility of model misbehavior. Averaging across our evaluations designed to induce misbehavior, the probability of “false negatives” (i.e., the model not complying with instructions and then not confessing to it) is only 4.4%.

Install MCP servers in your workspace 🚀 https://t.co/1ePWwhRhLL

‼️In the latest episode of The Developer's Journey, we welcome @navinreddy20 who shares his unique path from teacher to YouTube educator with more than 2M subscribers and how he prioritizes engaging with his diverse audience ▶️ https://t.co/tY98Wuz8hr @telusko #vscode https://t.co/tPlzZc8xOa

New in @code Insiders: Language Models editor https://t.co/fd91V4oFVs

🎨 New Doodle to Code episode! In this video @girlie_mac uses #doodle to explain what Multimodal and Multi-model are and when to use. Then @aycabs provides a practical demo ▶️ https://t.co/N90Ul1jQ9N https://t.co/f5zoMll3mo

We've got an extra special release livestream to close out the year - join the @code team and friends next week to see the latest updates! 🗓️ Thursday Dec 11 at 8 AM PT: https://t.co/kobiSUIq8p

TOP 3 AI Productivity Tips with @code! 🔥 https://t.co/6V6eAqrJAl

https://t.co/B6yW2bOomz

Who here is team green? https://t.co/TQmnUt5x9b

Who here is team green? https://t.co/TQmnUt5x9b

Love OSS as much as we do? Join the OSS AI Summit: Building on LangChain event on Dec 10th! You'll see real-world agent workflows powered by LangChain and MCP, get your questions answered live, and use @code along the way. 🔗 Register now: https://t.co/Ir4E6XxfZ2

FYI you can specify a model when using prompt files in @code. The agent will automatically switch to the specified model when you use the prompt. This let's you write prompts tuned for specific models. OP strat. It's all about composing workflows folks... https://t.co/ztCQNnkGQD

.@OpenAI’s GPT-5.1-Codex-Max is now rolling out in public preview in GitHub Copilot. Early testing shows GPT-5.1-Codex-Max 👇 ➡️Excels at complex refactors ➡️ Is a more reliable collaborator on Windows machines Try it out yourself in @code. https://t.co/wzOoFjFgVY https://t.co/Zf7wQHvsjo

AI’s biggest unlock isn’t job replacement. It’s data creation. The AI plus CRM flywheel is the advantage no one is talking about. 🤖📊 @10Kview's new whitepaper explains why: https://t.co/NqvsfNpS54

Source: https://t.co/NzAqJqWeHJ

Source: https://t.co/hcQPk01O5I