Sora 2 was releasedSora 2 was released
Sora 2 was the talk of the town. OpenAI releases Sora 2 - SOTA video and social network, @gmd shared, after it was properly pre-announced: OpenAI Is Preparing to Launch a Social App for AI-Generated Videos was shared by @0xbitcoiner right before that. You can get a Sora invite code, thanks @carter, who then went on to use it to create his first Sora video. The Verge thinks that it is a slippery slop, @jakoyoh629 shared.
More Sora 2:
- OpenAI’s New Video App Is Jaw-Dropping (for Better and Worse) by @ch0k1
- OpenAI's new Sora video generator to require copyright holders to opt out by @StillStackinAfterAllTheseYears (#1244866 by @RideandSmile)
- OpenAI launches new AI video app spun from copyrighted content by @Coinsreporter
More monetization, because you are the productMore monetization, because you are the product
Ad-ificial intelligence by @BlokchainB hits the nail right on the head, as Meta plans to sell targeted ads based on data in your AI chats, @optimism and @Kayzone (#1245486) shared. And this is such a great feature that you can't turn it off: @0xbitcoiner shared Meta won’t allow users to opt out of targeted ads based on AI chats.
On the upstream milking side, Music labels have figured out a way to charge AI companies by @lunin and @Scoresby shared that In the fight over AI, copyright is America’s competitive weapon. Also, Cloudflare is piloting an AI Index for all their customers, shared by @0xbitcoiner.
Conclusion: you will be forced to watch ads so that Zuck can pay the record companies.
Safety & SecuritySafety & Security
- First Malicious MCP in the Wild: A Postmark Backdoor That's Stealing Your Emails by @carter
- This streamer got deepfaked. Then the men showed up at her door. by @0xbitcoiner
Reports & ExplainersReports & Explainers
- Opti's Claude 4.5 Sonnet vibe coding" report" by @optimism
- Zero-Click Attacks: AI Agents and the Next Cybersecurity Challenge by @jakoyoh629
- Effective context engineering for AI agents by @carter
- How One AI Model Creates a Physical Intuition of Its Environment by @0xbitcoiner
- AI's Latest Buzzword: What Even Is Context Engineering? by @carter
- How to build a multilingual conversational AI voice assistant by @0xbitcoiner
- Why We Think by @carter
- Designing agentic loops by @deSign_r
- The Psychology Of Trust In AI: Measuring And Designing For User Confidence by @deSign_r
New ModelsNew Models
- MinerU2.5: A Decoupled Vision-Language Model for Document Parsing by @optimism
- GLM-4.6: Advanced Agentic, Reasoning and Coding Capabilities by @carter
- Experimental model release: DeepSeek-V3.2-Exp by @carter (#1243078 by @Cje95)
- Claude Sonnet 4.5 by @hn
ResearchResearch
- GDPval: Measuring the performance of our models on real-world tasks by @Scoresby
- Evaluating the Impact of AI on the Labor Market: Current State of Affairs by @0xbitcoiner
- I Created a New Kind of AI with No Neural Nets by @carter
- Unsexy AI Failures: The PDF That Broke ChatGPT by @Scoresby
- AI Can Now Do Expert-Level Work (Almost). 5 Surprising Findings by @RideandSmile
- LLM-Deflate: Extracting LLMs Into Datasets by @carter
- Efficient LLM Inference by @carter
- Video models are zero-shot learners and reasoners by @0xbitcoiner
- Sandboxing agents at the kernel level by @carter
- Claude_Sonnet-4.5 System Prompt - elder-plinius by @Scoresby
- CWM: An Open-Weights LLM for Research on Code Generation with World Models by @carter
- Just How Resilient Are Large Language Models? by @carter
- To AI or not to AI by @carter
OpinionOpinion
- Jeff Bezos hails AI boom as ‘good’ kind of bubble by @lunin
- iRobot Founder: Don’t Believe The (AI & Robotics) Hype! by @carter
- What industry will be affected by AI the most? by @nolem
- The Economics of AI: Dispelling Fears and Embracing Entrepreneurship by @RideandSmile
- Pope: AI can never replace the human touch of the doctor by @Kayzone
- AI trained for treachery becomes the perfect agent by @Scoresby
- Where’s the AI design renaissance? by @deSign_r
- Worshiping at the altar of AI by @Kayzone
- AI will happily design the wrong thing for you by @carter
- LLMs Are the Ultimate Demoware by @carter
Tools & integrationsTools & integrations
- Famed gamer creates working 5 million parameter ChatGPT AI model in Minecraft by @RideandSmile
- JPMorgan’s blueprint by to become the world’s first fully AI-powered megabank by @Diego
- SnapScalp: AI Tool to Recommend Market Trade Ideas for Intraday Scalpers by @carter
- Sculptor: The missing UI for Claude Code by @carter
- AI in the World of Ornithology: Using AI to Learn More About Birds by @RideandSmile
- AI Comes to Open Source eBook Reader Calibre by @RideandSmile
- Buy it in ChatGPT: Instant Checkout and the Agentic Commerce Protocol by @carter
- AI Leaders Outpace Laggards with Double the Revenue Growth and 40% Cost Savings by @Kayzone
- Dancing with AI: how next-gen game designers are taking the lead by @deSign_r
- Announcing fossabot: AI Agent for Strategic Dependency Updates by @carter
- ProofOfThought: LLM-based reasoning using Z3 theorem proving by @hn
- Cursor: Agent Autocomplete, Hooks, and Team Rules by @carter
- ChartDB Agent – Cursor for DB schema design by @hn
GovernmentGovernment
- Open Source Opportunity for AI Adoption in Africa, the Middle East, and Türkiye by @0xbitcoiner
- Germany is investing billions in AI - but who is making a profit? by @Kayzone
- Gavin Newsom signs first-in-nation AI safety law by @ch0k1 (#1243816 by @0xbitcoiner)
- The Demonization of DeepSeek: How NIST Turned Open Science into a Security Scare by @hn
Big Tech & MarketsBig Tech & Markets
- "Introducing the world's most powerful model" dosey doe by @k00b
- AI devs close to scraping bottom of data barrel by @0xbitcoiner
- Anthropic Researchers Run Into Trouble When New Model Realizes It's Being Tested by @south_korea_ln
- Putting ChatGPT on the Couch by @0xbitcoiner
- Circular Financing: Does Nvidia's $110B Bet Echo the Telecom Bubble? by @hn
- OpenAI is routing GPT-4o to safety models when it detects harmful activities by @0xbitcoiner
- AI Bubble Watch: Is it a Floater or a Popper? by @0xbitcoiner
- The good, the bad, and the future of AI agents by @jakoyoh629
- The Top Countries Driving ChatGPT Traffic in 2025 by @0xbitcoiner
- AI chipmaker Cerebras withdraws IPO by @ch0k1
- OpenAI Valuation Soars to $500 Billion, Topping Musk’s SpaceX by @ch0k1
- The AI Application Spending Report: Where Startup Dollars Really Go by @0xbitcoiner
- AI chip company Cerebras raises $1 billion in pre-IPO funding round by @Coinsreporter
- Helios AI Raises $4.7 Million and Launches Helios Horizon AI Supply Chain Tool by @Kayzone
I remember when Sora 1 was released and everyone we freaking out and posting all these amazing videos. But I remember trying it and not being able to quickly make anything good. And thus, it hasn't made much of an impact in my life, and I don't actually see it making a huge impact in society either, other than a few youtube channels dedicated to AI videos (like transforming cartoon characters to realistic looking people, for example). So either I'm just bad at prompting, or it still takes a lot of effort to get it to do what you want, well.
I wonder if OpenAI pays influencers to work on really nice Sora-generated videos, so that when they post about it and makes Sora look awesome, but the creators don't divulge how many hours of effort actually went into making the video look exactly the way they wanted.
Maybe @carter can comment on how long it took to make his demo vid
they seem to be about 2-3 min to generate depending on the duration
And you just vibe it - nothing special?
Great Analysis.