Exclusive: Deeper look into Hatch Agent from Meta
Meta’s unreleased Hatch materials point to a standalone agent platform with web, iOS, and Android apps, persistent project spaces, connectors, privacy controls, shareable agents, and browser or file-based task execution.
🗞 #meta @testingcatalog
Meta’s unreleased Hatch materials point to a standalone agent platform with web, iOS, and Android apps, persistent project spaces, connectors, privacy controls, shareable agents, and browser or file-based task execution.
🗞 #meta @testingcatalog
TestingCatalog AI News
Exclusive: Deeper look into Hatch Agent from Meta
Meta is testing Project Hatch AI Agent internally with new mobile apps, computer & browser use, goals support, spaces, customizable agent avatars, WhatsApp & Messenger integrations, and more!
🚨 AI News | TestingCatalog
Exclusive: Deeper look into Hatch Agent from Meta Meta’s unreleased Hatch materials point to a standalone agent platform with web, iOS, and Android apps, persistent project spaces, connectors, privacy controls, shareable agents, and browser or file-based…
BREAKING 🔥: "Project Hatch" from Meta will be a superapp with browser & computer use, /goals support, persistent cloud environments, spaces, and more!
Here is what we know so far 👀
> Hatch has been tested internally as a standalone web, iOS, and Android app.
> It supports browser & computer use with an approval flow, familiar from Codex or Claude Desktop.
> It supports long-running goals and task scheduling.
> Users will be able to customize "hatched" AI agents with different avatars.
> We will likely be able to create multiple AI agents, as we've seen in Grok Bot, with an option to share them.
> Hatch agents will have dedicated cloud environments, which could enable session persistence across devices.
> Users will be able to create Spaces. Memory and Voice Mode will also be supported.
> Messenger and WhatsApp Companions will let users communicate with Hatch via messaging apps.
* These details come on top of what has been reported earlier by Business Insider, based on the internal memo.
** It is possible that Meta will release only a subset of these features at the beginning.
Here is what we know so far 👀
> Hatch has been tested internally as a standalone web, iOS, and Android app.
> It supports browser & computer use with an approval flow, familiar from Codex or Claude Desktop.
> It supports long-running goals and task scheduling.
> Users will be able to customize "hatched" AI agents with different avatars.
> We will likely be able to create multiple AI agents, as we've seen in Grok Bot, with an option to share them.
> Hatch agents will have dedicated cloud environments, which could enable session persistence across devices.
> Users will be able to create Spaces. Memory and Voice Mode will also be supported.
> Messenger and WhatsApp Companions will let users communicate with Hatch via messaging apps.
* These details come on top of what has been reported earlier by Business Insider, based on the internal memo.
** It is possible that Meta will release only a subset of these features at the beginning.
OpenAI ❌ Cursor
OpenAI is ending its partnership with Cursor on November 12, following its acquisition by SpaceXAI.
AI tic-tac-toe (O vs X) 🤖
OpenAI is ending its partnership with Cursor on November 12, following its acquisition by SpaceXAI.
OpenAI models serve near 5% of Cursor’s traffic. Anthropic also responded with a commitment to remain being a trusted partner with Cursor and continue increasing compute capacity to support it.
AI tic-tac-toe (O vs X) 🤖
Tencent released open-source Hy4 preview model
Tencent unveiled Hy4 preview, a 770B-parameter model with a 1M-token context, built for software, office, gaming, and research work. It outperformed GLM 5.3 and Kimi K3 in internal tests and is now open source.
🗞 #ai @testingcatalog
Tencent unveiled Hy4 preview, a 770B-parameter model with a 1M-token context, built for software, office, gaming, and research work. It outperformed GLM 5.3 and Kimi K3 in internal tests and is now open source.
🗞 #ai @testingcatalog
TestingCatalog AI News
Tencent released open-source Hy4 preview model
Hy4 preview is now open-source and available via Tencent Cloud, OpenRouter, Tencent products and public repositories.
❤4👍1👀1
First outputs from GPT-6 "Astra" model from OpenAI
OpenAI appears to be nearing Astra’s release as internal testing expands around “mozaik-alpha-fdm.” Early zero-shot samples suggest strong coding and visual creation, though deployment may depend on safety and government review.
🗞 #openai @testingcatalog
OpenAI appears to be nearing Astra’s release as internal testing expands around “mozaik-alpha-fdm.” Early zero-shot samples suggest strong coding and visual creation, though deployment may depend on safety and government review.
🗞 #openai @testingcatalog
TestingCatalog AI News
First outputs from GPT-6 "Astra" model from OpenAI
The latest surfaced outputs were reportedly generated zero-shot with Max effort, where Astra spends substantially longer reasoning than GPT-5.6 Sol.
👍4
🚨 AI News | TestingCatalog
First outputs from GPT-6 "Astra" model from OpenAI OpenAI appears to be nearing Astra’s release as internal testing expands around “mozaik-alpha-fdm.” Early zero-shot samples suggest strong coding and visual creation, though deployment may depend on safety…
Media is too big
VIEW IN TELEGRAM
ASTRA 🔥: First outputs of the upcoming unreleased model from OpenAI have arrived, and these results are stunning!
> Internally named "mozaik-alpha-fdm", rumored to be a checkpoint of the "Astra" model, mentioned by OpenAI earlier as a substantial step-forward improvement.
Voxel Castle from the zAI Discord 👀
> Internally named "mozaik-alpha-fdm", rumored to be a checkpoint of the "Astra" model, mentioned by OpenAI earlier as a substantial step-forward improvement.
Voxel Castle from the zAI Discord 👀
❤6👍5
ANTHROPIC 🔥: Claude limits will be permanently increased by 25% starting from September 14.
What does this all mean?
Currently active, temporary 50% increase will end on September 14.
Meaning that your usage limit will be cut by ~17% after September 14 in comparison to what you have now.
Applies for Pro, Max, Team, and seat-based Enterprise plans.
What does this all mean?
Currently active, temporary 50% increase will end on September 14.
Meaning that your usage limit will be cut by ~17% after September 14 in comparison to what you have now.
😁9
OPENAI 🔥: Codex and ChatGPT Work usage limits are being reset once again.
Besides that, OpenAI is fixing various usage consumption issues so your overall usage limit will become higher than before.
K.O. 👀
Besides that, OpenAI is fixing various usage consumption issues so your overall usage limit will become higher than before.
Depending on how you use Codex, you should see your usage go between 10% and 50% further than before.
K.O. 👀
👍15❤5👀3
Google prepares Interactive Reports for Gemini Notebook
Google is preparing Gemini Notebook Interactive Reports, a feature for creating research-focused reports with embedded Studio content. Release messaging suggests an upcoming rollout soon.
🗞 #google @testingcatalog
Google is preparing Gemini Notebook Interactive Reports, a feature for creating research-focused reports with embedded Studio content. Release messaging suggests an upcoming rollout soon.
🗞 #google @testingcatalog
TestingCatalog AI News
Google prepares Interactive Reports for Gemini Notebook
Gemini Notebook is about to get the Interactive Reports feature - "An interactive report with embedded Studio content."
❤3
🚨 AI News | TestingCatalog
Google prepares Interactive Reports for Gemini Notebook Google is preparing Gemini Notebook Interactive Reports, a feature for creating research-focused reports with embedded Studio content. Release messaging suggests an upcoming rollout soon. 🗞 #google…
Media is too big
VIEW IN TELEGRAM
GOOGLE 🔥: Interactive Reports are being prepared for the upcoming release in Gemini Notebook.
> New: You can now create interactive reports.
> An interactive report with embedded Studio content.
Users can customize these reports by providing a custom prompt and selecting a language.
Based on the suggested prompt, this feature lets users generate interactive user research reports, and complex visualizations.
* Not available to the public yet.
> New: You can now create interactive reports.
> An interactive report with embedded Studio content.
Users can customize these reports by providing a custom prompt and selecting a language.
Based on the suggested prompt, this feature lets users generate interactive user research reports, and complex visualizations.
* Not available to the public yet.
Anthropic is working on an option to run commands in a local sandbox for Claude Code desktop.
> Runs commands from Claude Code in an isolated sandbox. Takes effect for new sessions.
A strict sandbox mode is available too.
> Blocks commands that can’t run inside the sandbox (ssh, for example) instead of running them outside it or asking you. Applies to every Claude Code session on this computer, including the terminal, unless a project’s own settings override it.
> Runs commands from Claude Code in an isolated sandbox. Takes effect for new sessions.
A strict sandbox mode is available too.
> Blocks commands that can’t run inside the sandbox (ssh, for example) instead of running them outside it or asking you. Applies to every Claude Code session on this computer, including the terminal, unless a project’s own settings override it.
❤4👍4
Perplexity prepares Hybrid mode for Computer on Mac
Perplexity is preparing Hybrid Mode for Mac, combining cloud orchestration with local inference for lighter tasks, lower credit use, and added privacy. A Privacy Gate may screen sensitive data before any cloud transfer.
🗞 #perplexity @testingcatalog
Perplexity is preparing Hybrid Mode for Mac, combining cloud orchestration with local inference for lighter tasks, lower credit use, and added privacy. A Privacy Gate may screen sensitive data before any cloud transfer.
🗞 #perplexity @testingcatalog
TestingCatalog AI News
Perplexity prepares Hybrid mode for Computer on Mac
What we know so far: Hybrid mode will delegate some subtasks to models hosted locally in order to save costs. Privacy Gate will evaluate data for PII before sending it to the cloud.
❤3🔥2
🚨 AI News | TestingCatalog
Perplexity prepares Hybrid mode for Computer on Mac Perplexity is preparing Hybrid Mode for Mac, combining cloud orchestration with local inference for lighter tasks, lower credit use, and added privacy. A Privacy Gate may screen sensitive data before any…
This media is not supported in your browser
VIEW IN TELEGRAM
PERPLEXITY 🔥: A new Hybrid mode is coming soon on Mac, letting local models handle some subtasks.
> Gemma 4 (for 16 GB machines), Qwen3.6, and Perplexity's own model (For 32 GB machines) are currently supported.
> "In Hybrid mode, Computer orchestrates in the cloud and delegates
suitable subtasks to a model on your Mac."
> "Local models run on your Mac's own memory and processor — the more capable your machine, the larger the models it can run."
> "Privacy Gate is a small model that runs entirely on your Mac. It inspects anything headed to the cloud for personal or sensitive information — and if it finds any, you decide how to handle it before anything is sent."
That would be huge 👀
> Gemma 4 (for 16 GB machines), Qwen3.6, and Perplexity's own model (For 32 GB machines) are currently supported.
> "In Hybrid mode, Computer orchestrates in the cloud and delegates
suitable subtasks to a model on your Mac."
> "Local models run on your Mac's own memory and processor — the more capable your machine, the larger the models it can run."
> "Privacy Gate is a small model that runs entirely on your Mac. It inspects anything headed to the cloud for personal or sensitive information — and if it finds any, you decide how to handle it before anything is sent."
That would be huge 👀
❤4👍2👀2🔥1