Ai
105 unrolled threads about ai, each one readable on a single page.
Alan Daitch@AlanDaitchGoogle acaba de lanzar esta BOMBA atómica de paper, quizás comparable con el que comenzó la revolución de IA. No exagero: explica por qué ChatGPT, Gemini,etc tienen el mismo problema: no pueden aprend...
SummaryGoogle publicó un paper revolutionary sobre "Nested Learning" que propone una solución elegante al problema de que los LLMs no pueden aprender después del entrenamiento. La clave es implementar un espectro continuo de memorias que operan a diferentes velocidades temporales (como el cerebro humano), en lugar del actual sistema binario de memoria rápida/congelada, permitiendo modelos que se adapten y mejoren continuamente durante su uso.Nov 23, 202514 tweets♥ 7.3K
Neuralink@neuralinkWe've been exploring how clinical trial participants can control a powered wheelchair — with their minds. Here's how we’re doing it 🧵
SummaryNeuralink has successfully enabled clinical trial participants to control a powered wheelchair using only their thoughts through a brain-computer interface with over 1,000 electrodes implanted in the motor cortex. A machine learning algorithm decodes neural activity to translate the participant's intention to move into cursor and wheelchair commands, allowing them to navigate real-world environments independently.Jul 23, 20266 tweets♥ 7.5K
Cynnie #jjstwt@gdnotfactsFor years I’ve been running an entirely AI generated roblox game, and nobody realised. Here's how I did it, and heres how you can too —🧵
SummaryA developer shares how they created a fully AI-generated Roblox game that went undetected for years, providing insights into the techniques and methods used for AI game development.Aug 6, 20262 tweets♥ 7.8K
Pablo Prompt@pablopromptDRAGON BALL Z – LIVE ACTION We’re on a roll lately… +8 MILLION views in 1 day with this video 😱 This trend is going super viral and a lot of people are making their own version. Hope you like mine...
SummaryA creator shares a tutorial on how to make viral Dragon Ball Z live-action videos using AI tools like Nano Banana Pro for image generation and Kling 2.5 Turbo for animation. The thread breaks down the process of generating realistic character images with specific prompts, editing them, and animating them into video sequences that have garnered millions of views.Dec 15, 20257 tweets♥ 7.5K
Google@GoogleToday, we’re introducing Personal Intelligence. With your permission, Gemini can now securely connect information from Google apps like @Gmail, @GooglePhotos, Search and @YouTube history with a singl...
SummaryGoogle has launched Personal Intelligence for Gemini, a new beta feature that lets the AI securely access your Gmail, Google Photos, Search, and YouTube data to provide more personalized recommendations and assistance. The feature is available starting today for Google AI Pro and Ultra subscribers in the US, with privacy controls allowing users to disconnect apps, adjust settings, or delete chat history at any time.Jan 14, 20268 tweets♥ 7.2K
gregorein@Gregoreinso... I audited Garry's website after he bragged about 37K LOC/day and a 72-day shipping streak. here's what 78,400 lines of AI slop code actually looks like in production. a single homepage load of...
SummaryA developer audited Garry Tan's website and found it loaded 6.42 MB across 169 requests for a simple newsletter-blog, containing AI-generated code bloat including test files shipped to production, duplicate assets, uncompressed 4MB PNGs, and broken image files. The critique highlights how AI code generation without proper review and shipping metrics without quality control led to severely bloated, wasteful production code.Mar 31, 202610 tweets♥ 7.6K
Michael Taiwo@AskMichaelTaiwoIf you want to get the most out of ChatGPT for writing, reviewing documents, learning, and planning. Then Copy and paste these prompts 🧵
SummaryA thread sharing 9 ChatGPT prompts designed to improve writing, document analysis, learning, and planning. The prompts cover extracting insights, finding weaknesses, improving readability, editing voice, fast learning, critical thinking, action planning, multi-perspective analysis, and knowledge synthesis.Aug 6, 202610 tweets♥ 5.9K
Nathan Calvin@_NathanCalvinOne Tuesday night, as my wife and I sat down for dinner, a sheriff’s deputy knocked on the door to serve me a subpoena from OpenAI. I held back on talking about it because I didn't want to distract f...
SummaryA California AI policy advocate describes being served a subpoena by OpenAI at his home, claiming the company used an unrelated lawsuit against Elon Musk as a pretext to intimidate critics of SB 53, a bill creating transparency and whistleblower protections for large AI companies. The author argues OpenAI's aggressive legal tactics and lobbying efforts contradict their stated nonprofit mission to benefit humanity.Oct 10, 202516 tweets♥ 6.2K
Google DeepMind@GoogleDeepMindThis is Gemini 3: our most intelligent model that helps you learn, build and plan anything. It comes with state-of-the-art reasoning capabilities, world-leading multimodal understanding, and enables...
SummaryGoogle announces Gemini 3, their most advanced AI model featuring state-of-the-art reasoning, multimodal capabilities, and improved coding abilities. Gemini 3 Pro is now rolling out globally and tops multiple AI benchmarks while offering more accurate, nuanced responses without filler.Nov 18, 20256 tweets♥ 6.4K
ClaudeDevs@ClaudeDevsOpus 5 is live in Claude Code and the Claude Platform today. A few things worth knowing: 🧵
SummaryClaude's new Opus 5 model is now available, achieving state-of-the-art performance on coding benchmarks with 43.3% on Frontier-Bench, excelling at complex multi-stage coding projects and offering both high-effort and fast modes for different use cases.Jul 24, 20267 tweets♥ 7.4KGoogle DeepMind@GoogleDeepMindSIMA 2 is our most capable AI agent for virtual 3D worlds. 👾🌐 Powered by Gemini, it goes beyond following basic instructions to think, understand, and take actions in interactive environments – mea...
SummaryGoogle announced SIMA 2, an advanced AI agent powered by Gemini that can understand and act in 3D virtual worlds with improved reasoning, generalization to unseen environments, and self-teaching capabilities. The system represents progress toward more capable AI agents with potential applications in robotics and real-world AGI development.Nov 13, 20255 tweets♥ 6.4K
Ilir Aliu@IlirAliu_You need $100 Billion to solve self-driving? Wrong. A small team of engineers in San Diego is doing it with a smartphone chip and a few cameras: ... for $999.🧵
SummaryGeorge Hotz's comma_ai is revolutionizing self-driving cars with a $999 device using smartphone chips and end-to-end learning, proving you don't need $100 billion and massive test fleets. The company has 20,000+ active users, runs profitably, and represents an open-source alternative to proprietary solutions from Tesla, Waymo, and Cruise.Dec 7, 202510 tweets♥ 7.2K
Sundar Pichai@sundarpichai1/ Gemini 3 in action, a quick🧵 You can give Gemini 3 anything (images, pdfs, scribbles, etc) and it will create whatever you like: an image becomes a board game, a napkin sketch transformed into a...
SummaryGoogle announces Gemini 3, an advanced multimodal AI that can process images, PDFs, and other inputs to generate creative outputs like board games, websites, and interactive lessons. The update includes improved reasoning capabilities for tasks like sports performance analysis, dynamic visual search results, and a new Gemini Agent that can handle complex multi-step tasks like booking services and organizing emails.Nov 18, 20255 tweets♥ 6.7K
Black Forest Labs@bfl_aiIntroducing FLUX 3. One multi-modal model for Image, Video, Audio and Action-Prediction. Creations are truer to life in every kind of style. FLUX 3 Video is now available in early access (link below...
SummaryBlack Forest Labs introduced FLUX 3, a unified multimodal AI model capable of generating and editing images, videos, and audio, plus predicting actions for robotics applications. The model is now available in early access for video generation, with additional features like native audio generation and open-weight access planned for rollout in coming weeks.Jul 23, 20266 tweets♥ 6.3K
Natalie Winters@nataliegwintersEXCLUSIVE: Chinese Institutions Are Building AI Models Of American Voters—And Testing Political Messages On Them. Fudan used 171 MILLION X posts to create a million-account “voter pool.” A government...
SummaryChinese research institutions have built AI systems to model American voters using 171 million X posts and Census data, then tested political messages on synthetic voter personas. Government-linked teams have studied how various demographics respond to campaign messages on topics like immigration and taxes, with some research explicitly designed to inform Chinese policy toward the U.S.Aug 21, 20266 tweets♥ 3.8K
Osvaldo Pinali Doederlein@opinaliDaniel got important clarifications from NVIDIA. TLDR: the "DLSS5 skeptical" were right about *everything*. 1⃣ It's a 2D AI Filter. Input is only color buffer & motion vectors. The model doesn't...
SummaryNVIDIA clarified that DLSS5 is a 2D AI filter that only sees color and motion data, lacking access to geometry and lighting information. The technology relies on guessing scene properties from pixels alone, explaining why demo scenes featured simple lighting and why it struggles with fast motion and complex illumination.Mar 19, 20267 tweets♥ 6.2KGoogle@GoogleToday we’re bringing @Gmail into the Gemini era, making it a personal, proactive inbox assistant to help you manage your life, not just your messages. Explore the new features launching today, many of...
SummaryGmail is integrating Gemini AI features to help manage inboxes more efficiently, including AI-powered filtering, email writing assistance, suggested replies, and email chain summaries. New features are rolling out to all users at no cost, with some advanced capabilities available to Google AI Pro and Ultra subscribers.Jan 8, 20265 tweets♥ 5.7K
Sebastian Meitz@MeitzPL⁉️ Czy “Markę Polska” ma nam wygenerować sztuczna inteligencja? ⁉️ Czy Ministerstwo Sportu i Turystyki zapłaciło 200 000 złotych w ramach zadania publicznego za raport wygenerowany przez AI? ⁉️ Czy...
SummaryA Polish Twitter thread questions whether a government-funded report on Poland's brand was generated by AI, citing numerous inconsistencies including numbered footnotes out of sequence, contradictory data, geographical errors, missing sources, and language that appears machine-generated. The Ministry of Sports and Tourism allegedly paid 200,000 PLN for the report attributed to a government official.Aug 20, 202615 tweets♥ 4.8K
Mistral AI@MistralAIIntroducing the Mistral 3 family of models: Frontier intelligence at all sizes. Apache 2.0. Details in 🧵
SummaryMistral released the Mistral 3 family of open-source models under Apache 2.0 license, including smaller Ministral models (3B, 8B, 14B) and the larger frontier-class Mistral Large 3 MoE, each available in base, instruct, and reasoning versions.Dec 2, 20254 tweets♥ 5.3K
SpaceXAI@SpaceXAI