Skip to main content
Back

OpenAI Codex now controls Windows desktop + mobile remote access

OpenAI launches biodefense platform, xAI releases coding model at $1/M tokens, Alibaba ships one model for 11 robots, Cursor adds auto-review mode

June 1, 2026 6 min read

Hey,

AI agents are no longer just suggesting code. They're taking over your screen, running your apps, and doing it from your phone.

OpenAI's Codex clicks through your Windows desktop while you steer from your couch. xAI's grok-build opens the same agentic loop to any developer for $1/M tokens. The coding assistant era is quietly dying.

Meanwhile, Alibaba's Qwen-VLA runs 11 robot types from one model. The same pattern is now eating robotics too.
Together with · partner_image
Author photo
Today's Author
Lior Alexander. Founder of AlphaSignal and former ML Engineer. Previously built ML systems at Iguazio, Guesty, Enphase, Mila.
Summary

Read time: 6 min 26 sec

Top News

▸ OpenAI brings Codex computer use to Windows with mobile controls

Tiger Data

▸ You Shipped an AI Feature. Your Database Felt It.

Top News

▸ xAI releases agentic coding model via public API at $1 per million tokens

GitLab

▸ AI Agents Need Context. GitLab Orbit Delivers.

Top News

▸ OpenAI launches biodefense platform giving U.S. government access to frontier biology models

Signals

1. Cursor ships auto-review mode so coding agents need fewer approval prompts
2. Presented by FriendliAI: DeepSeek, GLM, Kimi, Nemotron: production inference for any open-source LLM
3. Open-source PyTorch guide trains a 13M parameter LLM on a single GPU
4. Alibaba's Tongyi Lab ships one robot model that controls 11 different robot bodies
5. Unsloth releases quantized GGUF build of Step-3.7-Flash, a 198B vision-language model
6. Best open-source models right now matched to your GPU's VRAM
Top News
OpenAI brings Codex computer use to Windows with mobile controls
4,021 Likes
OpenAI brings Codex computer use to Windows with m

Mac users had all the fun. Now Windows catches up. Codex computer use just landed on Windows, and it changes what an AI coding tool can actually do.

Before this, Codex could only suggest code. Now it can operate your desktop directly, seeing your screen, clicking buttons, and typing, just like you would. Think: reproducing a UI bug, testing a desktop app, or navigating a tool that has no API.

  • Codex can click through your Windows apps to test and debug visual flows
  • Trigger it with @Computer or @AppName in your prompt
  • Control everything remotely from the ChatGPT mobile app on iOS or Android
  • Review diffs, approve actions, and steer tasks from your phone while your PC does the work

One heads-up: on Windows, computer use runs in the foreground, so Codex takes over your screen while it works. Also, your screen contents are processed through OpenAI's servers, so keep that in mind for sensitive projects.

Get it from the Microsoft Store. Update ChatGPT mobile too.

Read more
forward →
Presented by Tiger Data
You Shipped an AI Feature. Your Database Felt It.

When you add AI to your app, the data profile changes overnight. Every prompt, response, and user interaction becomes a timestamped event.

Vanilla Postgres handles it until it doesn't.

TimescaleDB extends Postgres to stay fast as that data piles up.

Hypertables partition automatically as volume grows.

Hypercore compression cuts storage 10x.

Continuous aggregates keep your dashboards live without re-querying everything.

Get $1000 credit
partner with us →
Top News
xAI releases agentic coding model via public API at $1 per million tokens
2,809 Likes
xAI releases agentic coding model via public API a

xAI just opened up grok-build-0.1 via public API, and it's worth paying attention to.

Most AI models are built for chat. This one is built to actually do work. It can autonomously plan, write, refactor, and iterate on code across multi-step tasks without you holding its hand the whole time. Think less "coding assistant" and more "coding collaborator that keeps going."

Here's what you get under the hood:

  • 256K context window, large enough to load an entire mid-sized codebase into memory at once
  • Accepts text AND images, so you can feed it UI mockups or error screenshots
  • Built-in tool use and reasoning, always on
  • Priced at $1/M input, $2/M output tokens, with cached prompts dropping to $0.20/M

You can plug it into Cursor, Kilo Code, OpenCode, or hit it directly via the xAI API or OpenRouter. Grab an API key and start building.

Read more
forward →
Presented by GitLab
AI Agents Need Context. GitLab Orbit Delivers.

AI agents are only as good as the context they have.

GitLab Orbit gives them a live knowledge graph of your entire software lifecycle so they move faster, cost less and get it right.

See it live at GitLab Transcend on June 10:

• Live demos of GitLab 19 & Duo Agent Platform
• Hands-on workshops
• The Developer Show

Stream it live
partner with us →
Top News
OpenAI launches biodefense platform giving U.S. government access to frontier biology models
1,565 Likes
OpenAI launches biodefense platform giving U.S. go

OpenAI just launched Rosalind Biodefense, a program that puts its life sciences AI model directly into the hands of teams building pandemic and biosecurity tools.

So what exactly is GPT-Rosalind? It reasons about molecules, proteins, genes, and disease-relevant biology, helping with evidence synthesis, hypothesis generation, experiment design, and data analysis. Think of it as a very smart biology lab assistant.

Here's what the program actually lets you build:

  • Epidemiological modeling, early detection, screening, preparedness, and non-pharmaceutical interventions
  • Early-warning systems, outbreak-response planning, diagnostics, and medical countermeasure development
  • Protein engineering and vaccine development, with partners like Lawrence Livermore and Johns Hopkins already on board

The big deal: access to GPT-Rosalind is free for approved teams, with OpenAI covering all access costs. This is vetted access, not open to everyone. Applications are open worldwide through OpenAI's website.

Read more
forward →
Signals
1
Cursor ships auto-review mode so coding agents need fewer approval prompts
1,208 Likes
2
DeepSeek, GLM, Kimi, Nemotron: production inference for any open-source LLM
Presented by FriendliAI
3
Open-source PyTorch guide trains a 13M parameter LLM on a single GPU
2,818 Stars
4
Alibaba's Tongyi Lab ships one robot model that controls 11 different robot bodies
417 Likes
5
Unsloth releases quantized GGUF build of Step-3.7-Flash, a 198B vision-language model
6,384 Downloads
6
Best open-source models right now matched to your GPU's VRAM
1,130 Likes
At Alpha Signal, our mission is to build a sharp, engaged community focused on AI, machine learning, and cutting-edge language models, helping over 200,000 developers stay informed and ahead. We're passionate about curating the best in AI, from top research and trending technical blogs to expert insights and tailored job opportunities. We keep you connected to the breakthroughs and discussions that matter, so you can stay in the loop without endless searching. We also work closely with partners who value the future of AI, including employers and advertisers who want to reach an audience as passionate about AI as we are.

Our partnerships are based on shared values of ethics, responsibility, and a commitment to building a better world through technology.Privacy is a priority at Alpha Signal. Our Privacy Policy clearly explains how we collect, store, and use your personal and non-personal information. By using our website, you accept these terms, which you can review on our website. This policy applies across all Alpha Signal pages, outlining your rights and how to contact us if you want to adjust the use of your information. We're based in the United States. By using our site, you agree to be governed by U.S. laws.
Looking to promote your company, product, service, or event to 250,000+ AI developers?
Work with us