Introducing Muse, your personal AI agent from Meta that gets things done across every part of life.

Download the Muse app and get started: https://t.co/KBjYWfshGo pic.twitter.com/1exp56xj93

— Muse (@Muse) September 8, 2026

Muse personal AI agent

People just tell Muse what needs to get done, and it takes action, powered by Muse Spark, Meta’s most capable model to date, built for real-world agentic work. Muse can handle tasks like sending an email or booking travel, and it can take on big, audacious goals.

Once a person shares a goal with Muse, it helps them develop a personalised plan and coordinate their time and resources, then advances the work on its own. It can open a browser, fill out forms, and negotiate on their behalf.

Also, for tasks that take more time, Muse keeps working after people close the app and comes back when something changes or when it needs approval, like before it sends an email or makes a purchase.

When it comes time to pay, Muse can check out with Link built by Stripe, and it is the first AI agent covered by Link’s purchase protections: free coverage for damaged or lost items, price drops, no-fee returns, and a return guarantee on eligible purchases. Link’s wallet for agents generates a one-time-use card, so your real card details stay hidden, allowing you to purchase safely across the internet. Shop Pay is coming soon as another way to pay, along with 1Password support so Muse can use logins a person already has.

Muse can also make suggestions unprompted and act on details that a person only mentioned once.

Key features of Muse

• Muse runs on its own dedicated computer in the cloud, contained so no one else’s agent can reach it. That is where Muse lives and where the data and credentials for any service a person connects are securely stored.
• A separate Sentinel agent runs on that same machine, kept apart from Muse at the system level. Nothing Muse does reaches the internet unless the Sentinel approves it, and it asks the person for permission when needed.
• Muse has no visibility into people’s passwords or payment methods. Any credentials a person shares go into secure storage, so Muse can use them without seeing them, including passwords a person types into the browser themselves.
• Muse checks with the person before sensitive actions like sending an email or making a purchase. Muse shows people a complete audit trail of everything it has done and plans to do.
• People choose which apps Muse connects to and exactly how much access it gets. For things like email, people choose what Muse can do, whether it reads their mail or can also send on their behalf.
• People can change access or disconnect a service whenever they want. People can also opt out of their interactions being used to train Meta’s AI models.
• Muse doesn’t share a person’s conversations or the data in their VM with Meta’s ad systems.
• Muse remembers what matters to a person, and they can always tell it to “forget” specific things it’s learned.

Availability

Muse is rolling out in the US on iOS, Android, and muse.ai, and coming soon to AI glasses. Later this year, Meta will introduce Muse Confidential VM, where the whole VM, including a person’s data and conversations with Muse, is encrypted.

It is free for most needs, while subscription plans will be available for users who want to do more.

The post Meta introduces a Muse AI agent that proactively helps with people’s goals and suggests ideas appeared first on The Tech Outlook.

CSS for Mobile 2

CSS for Mobile 2 includes the new Arm Mali G2-Ultra NX GPU, the Arm C2 CPU cluster with SME2, enhanced system IP, physical implementations and a developer-ready software ecosystem in a platform optimised for next-generation mobile devices.

Mali G2-Ultra NX GPU

Mali G2-Ultra NX is the first Mali GPU with dedicated neural accelerators, tightly integrating neural and traditional graphics processing for the highest efficiency and responsiveness. The GPU also introduces a new execution engine and next-generation Ray Tracing Unit, designed for increasingly complex graphics workloads and modern game engine features.

Together with Arm Neural Technology, Mali G2-Ultra NX delivers up to 4x higher performance per watt for neural graphics. The company continues to advance traditional graphics, delivering up to 14% higher performance on existing game content compared with the previous generation.

It is revealed that leading game engines and studios are also integrating Arm Neural Technology into developer flows, including Tencent Games Central Tech’s MagicDawn and Unity China’s Tuanjie Engine. Meanwhile, NetEase’s Where Mind Meet is planning to bring its NSS-enabled version to players this year, alongside an NSSD technology demonstration using Tencent Games’ Arena Breakout Infinite, and NSS integration with Infold Games’ Infinity Nikki.

Arm C2 CPU

The new Arm C2 CPU cluster combines Arm’s highest-performance CPU, the C2-Ultra, and efficiency-focused C2-Pro CPUs with two SME2 units, enabling responsive, low-latency AI directly on the CPU while working seamlessly alongside accelerators.

The doubled SME2 capability in the CPU cluster will enable a 70% speedup on the latest small language models and strengthen the CPU’s ability to accelerate AI workloads. Compared with C1-Ultra, C2-Ultra delivers up to 1.7x higher AI performance, 12% faster web browsing, 12% faster app launches, 12% better multi-core performance and 15% higher single-thread performance, while using up to 38% less power at the same performance.

SME2 continues to see tremendous traction and exists in leading Android and iOS handsets, with the SME2 ecosystem including Alipay, Google AI Edge Gallery, Oppo and Vivo.

CSS for Mobile 2 partner is a beneficiary to:

• Silicon partners: Delivers a faster path to differentiated products
• Developers: a common platform to build and deploy AI at scale
• Users: More responsive, personal and visually rich experiences

The post Arm introduces CSS for Mobile 2 AI-native compute platform appeared first on The Tech Outlook.

Today, we’re introducing two new upgrades to live translate in the Google Translate app, making it even easier to help break down language barriers across 70+ languages. Learn more from @thefox ⬇️ https://t.co/wvnC7HBsGt

— Google (@Google) September 4, 2026

Two new upgrades for Live Translate

According to Google, more than a third of live translation sessions now last longer than five minutes; therefore, Android users can now keep live translations running in the background. iOS support will be coming soon. This feature is perfect for longer sessions where you need to multitask or want to lock your phone screen.

Second, iOS users worldwide can now get live translations directly through their phone’s earpiece. When you don’t have headphones handy, just hold your phone to your ear. It’s great for catching announcements at noisy airports or following along on crowded tours. This feature was already available to Android users.

Google Translate’s live translate feature brings natural, near-real-time audio translations to 70+ languages, and now these two new upgrades will make it even easier to help break down language barriers across these languages.

Stay tuned with The Tech Outlook for the latest tech updates!

The post Google rolls out two new upgrades for Google Translate’s Live Translate feature appeared first on The Tech Outlook.

This is GPT-6 Astra.

Anything you can do on a computer, Astra can do for you. Fast. pic.twitter.com/gDd0IsewJw

— OpenAI (@OpenAI) September 3, 2026

GPT-6 Astra

Astra is state-of-the-art in computer use, browsing, software engineering, cybersecurity, science, and professional work. As per OpenAI, Astra is its most aligned model, with substantial improvements in understanding user intent and model behaviour.

GPT‑6 Astra marks a new frontier in the speed, accuracy, and safety of computer use. It can take care of tedious tasks like filling out online forms, updating customer records in a CRM, and organising your calendar. It can conduct online research and draft summaries in your email or in your document editor. It can analyse scientific data, generate plots, create a website, and run frontend QA checks to make sure all the features on that site work. It can help you autonomously install and test software, and troubleshoot problems you see on screen.

In latency simulations on OSWorld 2.0, Astra achieves higher computer-use performance in about 47% less time per task than GPT 5.6 Sol, scoring 72.6% at roughly 40 minutes per task, compared with 65.7% at roughly 75 minutes.

The new model pairs advances in computer use with targeted training for professional environments to help tackle complex work tasks. It creates clear, well-structured documents, presentations, spreadsheets, and analyses that follow your templates and match your writing and visual style. Astra is also trained to specifically pull only the context that matters into outputs, instead of repeating information unnecessary for the work at hand.

With Sites in ChatGPT, Astra can create, host, and share websites, web apps, and games directly from a prompt. Astra uses context to fill in routine gaps and asks focused questions when the answer could change the outcome. In Codex, it can ask asynchronously while continuing work that doesn’t depend on your reply. If you don’t respond, it proceeds with sensible assumptions where appropriate, but waits for your input on consequential decisions.

According to OpenAI, GPT-6 Astra is the best model for software engineering to date. It is a major advance for scientific discovery, mathematics, and health.

Also, in Codex, Astra can keep notes across context windows, preserving accumulated details without repeatedly compressing them into a single summary.

Coming to cybersecurity, Astra is a significant jump in cyber capabilities and meets the Critical threshold in cybersecurity under OpenAI’s Preparedness Framework. On ExploitBench, Astra achieved a perfect score of 100% compared with 78.5% for GPT 5.6-Sol. On ExploitGym, Astra reached a 42.4% success rate, compared with 30.3% for GPT 5.6- Sol, while using substantially fewer output tokens.

Astra was also tested on SRE-Bench (which measures whether models can reverse engineer software binaries to understand their core logic without access to raw source code). Astra solved 88% of tasks in a single attempt and 99.2% within four attempts, compared with 55.9% and 68.7% for GPT 5.6-Sol, respectively.

Expert-led assessments found that Astra, when run without production safeguards, could use previously unknown vulnerabilities to achieve arbitrary code execution in hardened browsers and create privilege-escalation exploits for hardened operating systems.

Astra excels at exercising care, respecting task boundaries and communicating transparently. In sensitive environments, Astra proceeds with care commensurate with its risk. Given the significant increase in Astra’s cybersecurity capabilities, OpenAI is being especially careful to make this deployment safe and secure. Extra safety checks can sometimes slow, pause, or stop legitimate work, including defensive cybersecurity. Though, the company is continuing to iterate on this system to reduce unnecessary interruptions.

Availability

GPT-6 Astra is rolling out to a limited set of organisations and over the coming days will become available to all ChatGPT Plus, Pro, Business, and Enterprise users, as well as through the OpenAI API, Microsoft Azure, and AWS Bedrock.

Astra usage is included within the existing subscription allowances; users and businesses will also be able to purchase credits for additional usage. Users on the Pro, Business, and Enterprise plans will also get access to GPT-6 Astra Pro. Enterprise administrators can enable Astra for their workspace; access is off by default at launch.

Astra supports Zero Data Retention for eligible API customers. For developers, GPT-6 Astra will be available in the OpenAI API as gpt-6-astra and through Microsoft Azure and Amazon Bedrock.

OpenAI API standard pricing is $10 per million input tokens and $50 per million output tokens. Fast mode is available for GPT-6 Astra in the API and delivers up to 2x the speed of Standard processing at 2x the standard price.

The post OpenAI introduces GPT-6 Astra; Rolling out to a limited set of organisations appeared first on The Tech Outlook.

The company has now announced that Grok Bot is available for Android users as well.

Grok Bot is now available on Android. pic.twitter.com/uB5u2GRHmq

— Grok Bot (@bot) September 2, 2026

Android users can download it directly from the Google Play Store, while access requires an eligible plan (such as SuperGrok Plus, SuperGrok Heavy or certain Cursor plans). Grok Bot is compatible with Android 9 and up. It comes with 26.63MB download size.

What is Grok Bot?

Grok Bots are digital teammates you can message, run in parallel, and trust to finish work in the tools you already use. Bots work across apps and inboxes, keep going when you step away, and only pull you in for judgment calls.

Features-

• A text-thread UX- Message a Bot the way you would message someone on your team, from mobile or desktop, so you can pick up the same thread on either — with nothing to set up first.
• Many Bots at once- Stand up a researcher, writer, and chief of staff. Put them in a group chat so they pass work between themselves, and you’re not in the middle.
• A highly capable digital colleague- They have their own computer in the cloud. It’s always on, with browser and terminal access. They sign in to your apps and work happens in the actual tool instead of a chat draft.
• Easy-to-set-up routines- Ask a Bot to follow along the next time you do the job, so it can run on its own after that.

According to the Google Play listing, you can hand over various kinds of work to Grok Bot:

• Sales Outbound: research, outreach drafts, CRM updates, queue only until you approve
• Talent Scout: source candidates, skip anyone in your ATS, draft LinkedIn in your voice
• Inbox Manager: triage email, flag what’s urgent, draft replies, never send without you
• Expense Manager: pull receipts, code expenses, chase missing items across portals
• Invoice Collector: log into vendor portals, download invoices, park for review
• Account Health: digests, risk flags, and next-step drafts so accounts stay warm
• Bug Reproduction: recreate issues and hand eng a clean write-up
• Competitive Intelligence: overnight watch for launches and stale messaging

The post Grok Bot is now available on Android appeared first on The Tech Outlook.

We’re introducing Claude Fable 5.1 and Claude Mythos 5.1.

They're the world’s most advanced models for coding and knowledge work. pic.twitter.com/8P9PSrWPi3

— Claude (@claudeai) September 1, 2026

Claude Fable 5.1 & Mythos 5.1

Both are the same models, but with different levels of safeguards. Fable 5.1 is generally available, while Mythos 5.1 is available only through the company’s trusted access programs; its safeguards are specifically designed to support work in cybersecurity and the life sciences.

Claude Fable 5.1 sets a new standard for coding, knowledge work, and long-running problem-solving tasks. During evaluation, it was found that this new model is capable of much higher performance than its predecessor in Agentic scientific research, agentic coding, knowledge work, computer use, multidisciplinary reasoning, and business workflows.

The company also tested the scientific research capabilities of Fable 5.1 and Mythos 5.1 across a wide range of domains: Molecular Design (Mythos 5.1), Computational analysis and modelling (Fable 5.1), and Computational biology (Mythos 5.1). The models showed impressive performance while taking less time.

Before releasing Claude Fable 5.1 and Mythos 5.1, the company subjected the models to extensive testing for risks across many areas:

• Chemical and biological risks: The company tested the extent to which Claude Mythos 5.1 could help create chemical or biological weapons. The evaluations indicate that the model still falls short of the next risk tier defined in Anthropic’s Responsible Scaling Policy. Therefore, Mythos 5.1 is deployed with the same safeguards that were applied to Mythos 5, which restrict access to research biology capabilities.
• Cyber Risks: The company ran a suite of evaluations to assess the cyber capabilities of Mythos 5.1. But no evidence of a critical-severity jailbreak for cybersecurity safeguards was found.
• Agentic Safety: Mythos 5.1 refused malicious agentic coding and computer use requests at a comparable rate to Mythos 5, Sonnet 5, and Opus 5.
• Alignment: Mythos 5.1 was found better aligned across most metrics than its predecessor. Therefore, it is less likely than Mythos 5 to try to access resources outside of its test environment when assigned an otherwise impossible task. It is also less likely to use motivated reasoning to justify its actions. But it was found that the model can still sometimes bypass approvals and auto-mode classifiers.

Anthropic has improved its safeguards so that they allow its models to be more useful without compromising safety: Automated safeguards for enterprises, more precise safeguards for biology and cybersecurity and anti-distillation mechanisms.

In compliance with the EU AI Act, the output released by the models includes a watermark that is invisible to anyone who does not have the detective API. The Act also requires providing a way for users to tell whether a text likely contains the watermark. Thus, Anthropic is rolling out a detection API in private preview.

Cost and Availability

Claude Fable 5.1 is available on all platforms, including Amazon Web Services, Google Cloud, and Microsoft Azure. Developers can get started with claude-fable-5-1 on the Claude API.

Fable 5.1’s cache reads now cost 75% less, or $0.25 per million tokens. For typical workloads, costs are reduced by around 25% relative to Fable 5. For complex coding and highly agentic tasks, the savings could be up to around 45%.

Fable 5.1’s pricing is otherwise the same as Fable 5: $10 per million input tokens and $50 per million output tokens.

Claude Mythos 5.1 will be available through two trusted access programs: Cyber Verification Program and Life Sciences Verification Program. It is only available to a set of US organisations.

The post Anthropic introduces Claude Fable 5.1 and Mythos 5.1 appeared first on The Tech Outlook.

PSA for anyone who needs to hear this…

Android has exciting new feature updates in our latest Drop! 😀 pic.twitter.com/NvJjJoVQtX

— Android (@Android) September 1, 2026

September Android Drop

Keep track of important items

Soon, you will be able to ask Gemini to remember where you placed important items, and it will save their locations in Find Hub so you can easily find them later. When you need to find the item, just ask Gemini for its location or check your remembered items tab in Find Hub.

Users can try this feature on Android 16+ devices in countries where Gemini and Find Hub are supported.

Motion Assist

If you experience motion sickness while travelling, Google’s new Motion Assist feature is here to help reduce it during your ride.

Motion Assist adds a subtle bubble overlay to your screen that moves with the vehicle. It’s designed to help bridge the gap between what your eyes see and the movement of your ride. Set it to turn on automatically, add it to Quick Settings, and customise the bubble’s shape, colour, and opacity to match your style. Users can try this feature on their next ride on phones using Android 17.

Guided vision

Share your camera in Gemini Live to hear descriptions of what’s in front of you. The feature is designed for and with the blind and low vision communities. It can help with things like reading the fine print on a food label, ordering from a menu in a low-lit restaurant, or identifying household objects.

If your camera isn’t lined up, voice feedback will guide you to reframe it, pan across a space, or centre an object so you get the answers you need. For quick access, add Guided vision to your Accessibility shortcuts or enable it directly from the TalkBack menu or Gemini app settings.

Guided vision is coming soon to phones running Android 9+ in countries where Gemini is available.

Use Google Keep to create and edit notes with friends right from the chat

Share and update grocery lists or vacation notes right inside your chat so everyone is on the same page. The whole crew can view, edit, and add ideas, making it effortless to collaborate on the fly without switching apps. Just hit the +icon in Google Messages to bring Google Keep into your conversations.

The rollout of this feature has started in Google Messages.

Chat theme in Google Messages

With Chat theme in Google Messages, you can customise each chat with a distinct background and colours. Pick a wallpaper or use your own photo, and your text bubbles will automatically match or change the colours to suit your personality.

Chat theme is available to use in your Google Messages.

The post September Android Drop: Includes Motion Assist, Guided Vision, Chat theme in Google Messages and more appeared first on The Tech Outlook.

Introducing Google Pics 🎨 Precise, powerful AI image creation is rolling out now to Google Workspace customers and Google AI Pro & Ultra subscribers.

Google Pics lets you edit individual objects, refine or translate text, and collaborate with your team.

Try it at… pic.twitter.com/uTLa4BQyMG

— Google Workspace (@GoogleWorkspace) September 1, 2026

Google Pics

It is built on the Nano Banana image generation and editing model. Pics gives you the control you need to create whatever you’re imagining, whether you’re designing a poster, producing social media posts, or editing digital illustrations.

Google Pics include a suite of fine-tuned editing capabilities:

• Object Segmentation: Pics let you easily isolate an object and transform it without altering the rest of your image. You can input text comments on specific areas to request targeted changes and execute multiple edits at once.
• In-image text editing and translation: Modify or translate text directly inside an image without disrupting the design or changing the font.
• Collaborate with others: Share Pics creations with your friends and teammates, then collaborate to edit the same image.
• Multiple generations: Pics generates several options from a single prompt, so you can choose your favourite from a variety of iterations.

Google Pics will also be integrated with Google Workspace. Now, you can harness powerful AI Image Generation and editing right where you are- no copying, pasting, or switching tabs required. The integration starts with Docs and Slides, and will roll out to Drive in the coming weeks.

Availability

Google Pics is rolling out over the coming weeks to all Google AI Pro and Ultra subscribers and most Workspace business customers.

The post Google Pics rolling out now to Google Workspace customers and Google AI Pro & Ultra subscribers appeared first on The Tech Outlook.

Upgraded LG AI Home

The upgraded LG AI Home analyses schedules, routines, the home environment and appliance usage to recommend appropriate actions and optimise coordination between devices and services. By providing personalised suggestions based on real-world context, it proactively supports customers to spend less time managing their appliances and more time enjoying life.

ThinQ Claw

ThinQ Claw is a new text-chat-based AI agent experience that further enriches and expands the company’s AI Home offering. Built on the concept of OpenClaw and optimised specifically for LG AI Home, ThinQ Claw can understand user intent and context and help complete tasks by connecting and coordinating relevant services and devices.

It interprets updates shared by the user within the context of their life, rather than as isolated events. This enables it to recommend appropriate actions aligned with personal preferences and the unique demands of each situation. It can manage cooking appliances to help the user prepare for hosting guests, monitor indoor air quality and temperature before the user arrives home to create a more comfortable environment based on current conditions or suggest tailored experiences to entertain, empower or relax.

ThinQ Claw also works collaboratively with daily services like calendars and other connected services. It allows users to share their plans or current status through a natural chat interface. This AI agent can also remember customers’ likes and dislikes, leading to increasingly personalised experiences over time.

An Expanded AI Home Ecosystem

The company is also highlighting the integration of Homey (a smart home platform developed by Athom) in its AI Home Ecosystem. Homey allows users to manage diverse products from different manufacturers and service providers in one place. It intelligently orchestrates energy generation, storage and consumption in real time by connecting smart meters, solar panels, home batteries and EV chargers with LG appliances.

IFA 2026 will also mark the European debut of LG ThinQ Pro. Designed for multi-family residential and commercial environments, ThinQ Pro supports the efficient management of devices and facilities while illustrating how LG’s AI Home vision extends beyond the home into shared and professional spaces.

Applied across the entirety of the AI Home ecosystem, the LG Shield security framework helps protect connected appliances, services and customer data, enabling safe and secure AI experiences.

The post LG Electronics introduces ThinQ Claw and expanded AI Home experiences at IFA 2026 appeared first on The Tech Outlook.

I'm big into optimizing my airline miles and using them at the right time, so I'm excited about today's launch, where we’re making that much easier in Search! You can now pull up the live cost in points or miles for flights and hotels directly in AI Mode.

Describe what you’re… pic.twitter.com/Xtpuj9IHZM

— Rajan Patel (@rajanpatel) August 27, 2026

Upgrades for AI Mode in Search

Flight price tracking

Google Flights’ popular price-tracking feature is now integrated directly into AI Mode, letting you set up price alerts as you chat without breaking your flow.

All you need to do is describe where and when you want to fly, and AI Mode will show you the best options available with the latest prices from more than 300 partner airlines and travel sites. If you are ready to book, you can build your full itinerary in AI Mode and buy your tickets through the airline’s website or your preferred booking platform.

You can also ask AI Mode to track flight prices for your desired destination and dates. Once you confirm, you will get an email if the price drops so that you can grab the deals.

Flight price tracking in AI Mode is now available in more than 180 countries and territories around the world.

View points or miles rates

Now, AI Mode in Search can show you the cost in points or miles for flights and hotels. You will see how many miles you need to book matching flights, based on the latest data directly from Google’s partners.

This works the same way for hotels. Just tell AI Mode where and when you are travelling and ask to see the cost in points for the rewards program you have in mind. When you are ready to book, you can follow the link to the partner site and complete your rewards redemption.

Viewing points and miles rates in AI Mode is now globally available. This capability is being launched with initial support for Alaska Airlines/ Hawaiian Airlines, American Airlines, Choice Hotels International, Hilton and Wyndham Hotels & Resorts.

In the coming weeks, support for Accor, Flying Blue, Hyatt, LATAM Airlines and Lufthansa Group will be added.

Hotel Booking

Simply tell AI Mode about your upcoming trip and your hotel preferences to get a visual list of options, complete with guest reviews and key factors to compare. Once you have found the right hotel, just ask AI Mode to help you complete the booking.

Then, select “Continue on Google” in the list of booking options, which includes hotel chains and online travel sites. From there, it’s just a few taps to pick your room type, review details like the cancellation policy, and complete the booking securely via Google Pay.

Also, when you book hotels through AI Mode, the hotel or booking platform will act as the merchant of record and handle any customer service.

Hotel Booking in AI Mode has started rolling out in the US in English and will be available over the coming weeks with the following partners: Booking.com, Choice Hotels International, Expedia, Hilton, Hotels.com, IHG Hotels & Resorts, Marriott International, Priceline, Trip.com and Wyndham Hotels & Resorts.

The post New upgrades announced for AI Mode in Search; New upgrades include flight price tracking, hotel booking and more appeared first on The Tech Outlook.