This media is not supported in your browser
VIEW IN TELEGRAM
π Google NotebookLM Gets Smarter!
β¨ Video Overviews now support 80 languages π
π§ Audio Overviews: longer, deeper summaries (not just highlights)
π Rolling out worldwide this week
β¨ Video Overviews now support 80 languages π
π§ Audio Overviews: longer, deeper summaries (not just highlights)
π Rolling out worldwide this week
This media is not supported in your browser
VIEW IN TELEGRAM
π¨ Big AI Update π¨
Anthropic just launched Claude for Chrome your AI co-pilot that lives inside the browser! ππ€
π It can click buttons, fill forms, manage emails, even schedule tasks while you surf.
β‘οΈ Limited pilot: only 1,000 Claude Max users get access first (with a waitlist).
π‘οΈ Comes with built-in safety checks like site permissions & confirmations.
π More features & wider rollout coming soon!
Anthropic just launched Claude for Chrome your AI co-pilot that lives inside the browser! ππ€
π It can click buttons, fill forms, manage emails, even schedule tasks while you surf.
β‘οΈ Limited pilot: only 1,000 Claude Max users get access first (with a waitlist).
π‘οΈ Comes with built-in safety checks like site permissions & confirmations.
π More features & wider rollout coming soon!
π AI Revenues Are Skyrocketing!
OpenAI, Anthropic, and 16 other AI-native startups are now generating a combined $18.5B in annualized revenue.
π Key Highlights:
π OpenAI + Anthropic = 88% of the total
π° Revenue surged massively in the past year (Aug β23 β Jul β25)
π Other rising players: Midjourney, Perplexity, Replit, Synthesia, ElevenLabs, Runway, Jasper, Cohere & more
π AI is no longer βfuture techβ itβs one of the biggest business engines of today.
OpenAI, Anthropic, and 16 other AI-native startups are now generating a combined $18.5B in annualized revenue.
π Key Highlights:
π OpenAI + Anthropic = 88% of the total
π° Revenue surged massively in the past year (Aug β23 β Jul β25)
π Other rising players: Midjourney, Perplexity, Replit, Synthesia, ElevenLabs, Runway, Jasper, Cohere & more
π AI is no longer βfuture techβ itβs one of the biggest business engines of today.
π¨ AI & Jobs Update
Ex-OpenAI researcher Miles Brundage says that with AI-driven growth, a $10,000/month UBI could be possible in the future. πΈπ€―
π Right now, $1K/month looks realistic, but AI could make much bigger payouts sustainable.
β‘ His point: as AI reshapes the job market, UBI may become a necessity, not just an experiment.
Ex-OpenAI researcher Miles Brundage says that with AI-driven growth, a $10,000/month UBI could be possible in the future. πΈπ€―
π Right now, $1K/month looks realistic, but AI could make much bigger payouts sustainable.
β‘ His point: as AI reshapes the job market, UBI may become a necessity, not just an experiment.
π€£1
This media is not supported in your browser
VIEW IN TELEGRAM
Introducing v2 of our SFX model.
Generate any sound effect directly from a prompt, via UI or API.
β’ Higher SFX quality
β’ New: seamlessly looping sound effects
β’ Maximum duration extended from 22s to 30s
β’ Sample rate increased from 44.1kHz to 48kHz
β’ Refreshed SFX Library
Generate any sound effect directly from a prompt, via UI or API.
β’ Higher SFX quality
β’ New: seamlessly looping sound effects
β’ Maximum duration extended from 22s to 30s
β’ Sample rate increased from 44.1kHz to 48kHz
β’ Refreshed SFX Library
This media is not supported in your browser
VIEW IN TELEGRAM
π¨Rolling out NEW audio overview formats:
(Default) Deep Dive: a thorough examination of your sources
Brief: 1-2 minute, bite-sized overviews
Critique: an expert review, offering constructive feedback on your material
Debate: a thoughtful debate between two hosts
(Default) Deep Dive: a thorough examination of your sources
Brief: 1-2 minute, bite-sized overviews
Critique: an expert review, offering constructive feedback on your material
Debate: a thoughtful debate between two hosts
This media is not supported in your browser
VIEW IN TELEGRAM
π Navigate PDFs Effortlessly with Grok!
π Highlight any section and tap "Explain" to instantly understand it.
π¬ Or click "Quote" to ask specific questions about the text.
Make reading PDFs smarter and faster with Grok! πβ¨
π Highlight any section and tap "Explain" to instantly understand it.
π¬ Or click "Quote" to ask specific questions about the text.
Make reading PDFs smarter and faster with Grok! πβ¨
π° Amazon Music introduces AI-powered βWeekly Vibeβ playlists
Amazon Music has launched a new feature called Weekly Vibe, an AI-generated playlist that updates every Monday.
πΉ The playlist is tailored to each userβs mood, listening history, and evolving taste.
πΉ It combines familiar tracks with new music discoveries.
πΉ Available under Library β Made for You.
πΉ Rolling out in the U.S. on iOS and Android for all subscribers.
Amazon Music has launched a new feature called Weekly Vibe, an AI-generated playlist that updates every Monday.
πΉ The playlist is tailored to each userβs mood, listening history, and evolving taste.
πΉ It combines familiar tracks with new music discoveries.
πΉ Available under Library β Made for You.
πΉ Rolling out in the U.S. on iOS and Android for all subscribers.
π¨ Big AI Shake-Up
Microsoft is cutting back reliance on OpenAI π
π The company will start using Anthropicβs Claude Sonnet 4 in Office apps (Word, Excel, PowerPoint, Outlook).
π Tests showed Claude beats GPT in some tasks (Excel automation, sleek PowerPoint slides).
π Access will come via AWS, since Amazon holds a big stake in Anthropic.
Microsoft says the OpenAI partnership stays, but itβs clearly moving to a multi-AI strategy.
Microsoft is cutting back reliance on OpenAI π
π The company will start using Anthropicβs Claude Sonnet 4 in Office apps (Word, Excel, PowerPoint, Outlook).
π Tests showed Claude beats GPT in some tasks (Excel automation, sleek PowerPoint slides).
π Access will come via AWS, since Amazon holds a big stake in Anthropic.
Microsoft says the OpenAI partnership stays, but itβs clearly moving to a multi-AI strategy.
π¨ OpenAI Upgrades Codex!
π» Meet GPT-5-Codex β the new AI coding powerhouse!
β‘ Smarter, faster, and can spend seconds to hours thinking depending on task complexity.
π Build projects, add features, debug, and review code like a pro.
π Available now on terminal, IDEs, GitHub & ChatGPT for Plus/Pro/Business users!
π» Meet GPT-5-Codex β the new AI coding powerhouse!
β‘ Smarter, faster, and can spend seconds to hours thinking depending on task complexity.
π Build projects, add features, debug, and review code like a pro.
π Available now on terminal, IDEs, GitHub & ChatGPT for Plus/Pro/Business users!
π YouTube adds Generative AI tools for Shorts creators!
β¨ Whatβs new:
β’ Veo 3 Fast β text-to-video with sound + faster rendering
β’ Animate still images with motion from other clips
β’ Fresh styles (pop art, origami & more)
β’ Edit with AI β auto-highlights, transitions, music & voice-over
β’ Speech-to-Song β turn spoken lines into songs πΆ
π Rolling out in US, UK, Canada, Australia, NZ β more soon.
β¨ Whatβs new:
β’ Veo 3 Fast β text-to-video with sound + faster rendering
β’ Animate still images with motion from other clips
β’ Fresh styles (pop art, origami & more)
β’ Edit with AI β auto-highlights, transitions, music & voice-over
β’ Speech-to-Song β turn spoken lines into songs πΆ
π Rolling out in US, UK, Canada, Australia, NZ β more soon.
This media is not supported in your browser
VIEW IN TELEGRAM
π° Product Update: Control GPT-5 Thinking Speed
Starting today, Plus, Pro, and Business users can adjust how long GPT-5 thinks β choosing the pace that fits the moment.
Hereβs whatβs new:
Standard β Balanced speed & intelligence (new default)
Extended β Deeper thinking (previous default for Plus)
Light β Snappiest replies (Pro only)
Heavy β Deepest reasoning (Pro only)
βοΈ How it works:
Select GPT-5 with Thinking in ChatGPT on web and toggle the thinking time in the message composer. Your choice will stay in place for future chats until you change it.
Starting today, Plus, Pro, and Business users can adjust how long GPT-5 thinks β choosing the pace that fits the moment.
Hereβs whatβs new:
Standard β Balanced speed & intelligence (new default)
Extended β Deeper thinking (previous default for Plus)
Light β Snappiest replies (Pro only)
Heavy β Deepest reasoning (Pro only)
βοΈ How it works:
Select GPT-5 with Thinking in ChatGPT on web and toggle the thinking time in the message composer. Your choice will stay in place for future chats until you change it.
π Google Gemini Comes to Chrome!
Google has rolled out Gemini in Chrome for U.S. users! πΊπΈ
Now your browser can do more than just browse β it can act.
β¨ Key Features:
AI can help with booking, shopping, navigating websites
Retrieve previously visited pages easily
Integrates with Google Calendar, Maps, YouTube
Suggests contextual questions for smarter browsing
π Rolling out this month in English for the U.S., with more countries coming soon!
Google has rolled out Gemini in Chrome for U.S. users! πΊπΈ
Now your browser can do more than just browse β it can act.
β¨ Key Features:
AI can help with booking, shopping, navigating websites
Retrieve previously visited pages easily
Integrates with Google Calendar, Maps, YouTube
Suggests contextual questions for smarter browsing
π Rolling out this month in English for the U.S., with more countries coming soon!
π Qwen3-Omni is here β the first truly omni-modal AI unifying text, image, audio & video in one seamless model. No more modality trade-offs!
π Benchmark Leader: #1 on 22/36 audio & AV tasks
π Supports: 119 languages (text) | 19 (speech in) | 10 (speech out)
β‘ Ultra-fast: 211ms latency | π§ Handles 30-min audio effortlessly
π¨ Fully customizable with system prompts
π Tool calling built-in
π€ Plus: Open-source Captioner with low hallucination
β¨ Whatβs Open-Sourced?
Qwen3-Omni-30B-A3B-Instruct
Qwen3-Omni-30B-A3B-Thinking
Qwen3-Omni-30B-A3B-Captioner
π Explore Now:
π¬ Qwen Chat
π» GitHub
π€ Hugging Face
π€ ModelScope
π¬ Demo
β‘ The future of omni-modal AI starts here!
π Benchmark Leader: #1 on 22/36 audio & AV tasks
π Supports: 119 languages (text) | 19 (speech in) | 10 (speech out)
β‘ Ultra-fast: 211ms latency | π§ Handles 30-min audio effortlessly
π¨ Fully customizable with system prompts
π Tool calling built-in
π€ Plus: Open-source Captioner with low hallucination
β¨ Whatβs Open-Sourced?
Qwen3-Omni-30B-A3B-Instruct
Qwen3-Omni-30B-A3B-Thinking
Qwen3-Omni-30B-A3B-Captioner
π Explore Now:
π¬ Qwen Chat
π» GitHub
π€ Hugging Face
π€ ModelScope
π¬ Demo
β‘ The future of omni-modal AI starts here!
β¨ Google Photos gets AI editing by voice & text!
Now on Android, you can edit pics in Google Photos just by talking or typing to AI π
πΌοΈ Say things like:
βRemove the person in the backgroundβ
βBrighten the skyβ
βMake it more vibrantβ
β‘ Powered by Gemini AI, it handles edits, tweaks & even complex object removal.
π Available now:
πΊπΈ U.S. only, in English, 18+ users
Now on Android, you can edit pics in Google Photos just by talking or typing to AI π
πΌοΈ Say things like:
βRemove the person in the backgroundβ
βBrighten the skyβ
βMake it more vibrantβ
β‘ Powered by Gemini AI, it handles edits, tweaks & even complex object removal.
π Available now:
πΊπΈ U.S. only, in English, 18+ users
Epoch Research claims:
GPT-5 used less pre-training compute than GPT-4.5, but GPT-6 will likely be trained on more again
that's because "reasoning" post-training improvements are slowing down,
and with larger GPU clusters available, scaling pre-training makes sense again.
openAI max-compute model + RL will be insane
GPT-5 used less pre-training compute than GPT-4.5, but GPT-6 will likely be trained on more again
that's because "reasoning" post-training improvements are slowing down,
and with larger GPU clusters available, scaling pre-training makes sense again.
openAI max-compute model + RL will be insane
This media is not supported in your browser
VIEW IN TELEGRAM
π¨ Big news for families! ChatGPT now lets parents link accounts with teens π
β Stronger teen safeguards
β Feature controls & limits
Rolling out on web today, mobile soon!
β Stronger teen safeguards
β Feature controls & limits
Rolling out on web today, mobile soon!
This media is not supported in your browser
VIEW IN TELEGRAM
Anthropic: A video of all versions of Claude, from the original to 4.5, trying to recreate claude.ai