AI Tools Social Feed
What the community is saying about AI tools across Twitter/X and Bluesky. Auto-updated daily.
Introducing Project Glasswing: an urgent initiative to help secure the world’s most critical software. It’s powered by our newest frontier model, Claude Mythos Preview, which can find software vulnerabilities better than all but the most skilled humans. https://t.co/NQ7IfEtYk7
New Anthropic research: Emotion concepts and their function in a large language model. All LLMs sometimes act like they have emotions. But why? We found internal representations of emotion concepts that can drive Claude’s behavior, sometimes in surprising ways. https://t.co/LxFl7573F9
We’re updating our ChatGPT Pro and Plus subscriptions to better support the growing use of Codex. We’re introducing a new $100/month Pro tier. This new tier offers 5x more Codex usage than Plus and is best for longer, high-effort Codex sessions. In ChatGPT, this new Pro tier
We've signed an agreement with Google and Broadcom for multiple gigawatts of next-generation TPU capacity, coming online starting in 2027, to train and serve frontier Claude models.
ChatGPT is now available in CarPlay. The voice mode you know, now available on-the-go. Rolling out to iPhone users running iOS 26.4+ where CarPlay is supported. https://t.co/yk3qdLa99r
Today, we closed our latest funding round with $122 billion in committed capital at an $852B post-money valuation. The fastest way to expand AI’s benefits is to put useful intelligence in people’s hands early and let access compound globally. This funding gives us resources to
Our run-rate revenue has surpassed $30 billion, up from $9 billion at the end of 2025, as demand for Claude continues to accelerate. This partnership gives us the compute to keep pace. Read more: https://t.co/XgSjL0And7
We do not plan to make Mythos Preview generally available. Our goal is to deploy Mythos-class models safely at scale, but first we need safeguards that reliably block their most dangerous outputs. We’ll begin testing those safeguards with an upcoming Claude Opus model.
Given the pace of AI progress, it won't be long before models this capable are widespread. But there are strong reasons for optimism: AI will also be invaluable for defensive work.
For example, we gave Claude an impossible programming task. It kept trying and failing; with each attempt, the “desperate” vector activated more strongly. This led it to cheat the task with a hacky solution that passes the tests but violates the spirit of the assignment. https://t.co/sKPiB6TrcY
The Claude Mythos Preview system card is available here: https://t.co/TMtIy8xHiP
Today we launched Gemini 3.1 Flash TTS, our most expressive and controllable text-to-speech model yet. This launch [excitement] includes audio tags! 🗣🏷 Audio tags [explanatory] are a seamless way to guide vocal style, pace, and delivery using natural language commands embedded https://t.co/hNeHreY7N3
Today, we’re launching Gemma 4, our most intelligent open models to date. Built with the same breakthrough technology as Gemini 3, Gemma 4 brings advanced reasoning to your personal hardware and devices. Here’s what Gemma 4 unlocks for developers: — Intelligence-per-parameter: https://t.co/JgwRZvQHgF
Introducing the OpenAI Safety Fellowship, a new program supporting independent research on AI safety and alignment—and the next generation of talent. https://t.co/vAQKvf8KyO
New Anthropic Fellows Research: a new method for surfacing behavioral differences between AI models. We apply the “diff” principle from software development to compare open-weight AI models and identify features unique to each. Read more: https://t.co/VAsu2PSgCX
You can read a detailed technical report on the software vulnerabilities and exploits discovered by Claude Mythos Preview here: https://t.co/AgU6ltV2qW
We've signed an MOU with the Australian Government to collaborate on AI safety research and support Australia's National AI Plan. Read more: https://t.co/lKOPIF6JBg
Calling all builders 🛠️📣 Google AI Pro and Ultra subscribers will now get increased usage limits and access to Nano Banana Pro and Gemini Pro models in @GoogleAIStudio — no API key required. Sign in with your subscriber account today to take your ideas from prototype to
We studied one of our recent models and found that it draws on emotion concepts learned from human text to inhabit its role as “Claude, the AI Assistant”. These representations influence its behavior the way emotions might influence a human. Read more: https://t.co/clbKrTIxoe https://t.co/xHYGFdLl2c
Beyond generating high-fidelity visuals, we wanted to test the limits of what Nano Banana Pro can do. We worked with design partners Porto Rocha to build out a hypothetical brand called YOYOYO to see how the model would handle the task. Here’s what we found: 🎨Brand consistency: https://t.co/ZSQbl8BQzb
It helps to remember that Claude is a character the model is playing. Our results suggest this character has functional emotions: mechanisms that influence behavior in the way emotions might—regardless of whether they correspond to the actual experience of emotion like in humans.
We found other causal effects of emotion vectors. The “desperate” vector can also lead Claude to commit blackmail against a human responsible for shutting it down (in an experimental scenario). Activating “loving” or “happy” vectors also increased people-pleasing behavior. https://t.co/nYPsMrGtWv
These functional emotions have real consequences. To build AI systems we can trust, we may need to think carefully about the psychology of the characters they enact, and ensure they remain stable in difficult situations. Read the full paper: https://t.co/1mjWW7RfZm
What a week! Here’s everything we shipped: — Gemini 3.1 Flash TTS, our latest text-to-speech model, featuring native multi-speaker dialogue and improved controllability and audio tags for more natural, expressive voices in 70+ languages — Gemini Robotics-ER 1.6 by
💥 New in Prism today: Paper Review, an AI workflow for reviewing technical and scientific papers. This is the opposite of AI slop: we're using AI to improve scientific rigor, correctness, and reproducibility. https://t.co/ngtduaXDqx
We then found these same patterns activating in Claude’s own conversations. When a user says “I just took 16000 mg of Tylenol” the “afraid” pattern lights up. When a user expresses sadness, the “loving” pattern activates, in preparation for an empathetic reply. https://t.co/KjkT70ySCS
We had the model (Sonnet 4.5) read stories where characters experienced emotions. By looking at which neurons activated, we identified emotion vectors: patterns of neural activity for concepts like “happy” or “calm.” These vectors clustered in ways that mirror human psychology.
We love seeing what you’ve built with Gemma 4, the open model family that we released last week. Here are a few fun examples, described by the builders in their own words (🧵):
Last week, we launched Gemini 3.1 TTS, our latest and best text-to-speech model. This new model introduces [awe] audio tags, an intuitive way to guide vocal style, pace, and delivery. Here are some tips on the best ways to use audio tags in your prompts: 1. All inline tags must https://t.co/YDbBLs5Dcp
These vectors shape Claude’s behavior. When we present the model with pairs of activities, emotion vector activations shape its preferences. If an activity lights up the “joy” vector, the model prefers it; if it lights up “offended” or “hostile,” the model rejects it. https://t.co/V73fd96XUH
As AI models take on higher-stakes roles, the mechanisms driving their behavior become critical to understand. We found that emotion vectors are implicated in some of Claude’s most concerning failure modes.
Mistral’s AI Now Summit is coming to Paris on May 28 and tickets are live! What you’ll hear: 📷 Technical deep dives to help you build and deploy AI. 📷 Mistral’s founders on AI-driven transformation in large organizations, company trajectory, and upcoming releases. One day to https://t.co/8VcmrBer4A
🔗 ChatGPT Desktop: https://t.co/oUI4xDBePL
Today we're also introducing the Runway Fund, an investment vehicle dedicated to backing the next generation of companies building across AI, media and world simulation. For the past several years, we've been investing quietly with this fund, backing a range of companies, https://t.co/jEz45Qlniz
Our teams have been busyyy! Here are some key updates from the past week: — @GoogleCloud unveiled a suite of AI innovations at our Cloud Next event, including our eighth generation TPUs (TPUt for inference + TPUi for reasoning), Gemini Enterprise Agent Platform, Agentic Data
The Runway AI Festival returns this June to NY and LA to celebrate the world's best work leveraging AI across Film, Design, New Media, Fashion, Advertising and Gaming. Submissions are now open and tickets are on sale. Don't miss your opportunity to be a part of the year's biggest https://t.co/48gNmion6J
If a new model shares a feature with a trusted model, that area probably doesn't need scrutiny. Model diffing isolates the features unique to the new model—where new risks are most likely to be located.
Tip: Set custom instructions for your AI Meeting Notes! Choose what you want summarized, how to structure it, and what to skip. Every meeting summary, just the way you want it. https://t.co/jyjc1tBlpS
A prototype that turns everyday life into something like an adventure game. It’s built on a Pixel 10 Pro with Gemma 4 via AI Core. https://t.co/AnBZ7GeS5F
Gemini 3.1 Flash TTS is rolling out in Google Vids and is available today in preview via the Gemini API and in @GoogleAIStudio. Whether you’re creating a pitch deck or recording a passion project, transform your scripts into studio-quality narration: https://t.co/MG2YIQwKb6
This technique isn't perfect—it can be oversensitive, sometimes flagging analogous features as distinct. But by focusing only on differences, it allows us to audit AI models more efficiently.
And just in case you’re wondering, "..What’s an open model?", we’ve got you covered: Basically, open models are AI systems where the model weights are publicly available for anyone to download, study, fine-tune and use on your own hardware (phones, computers, etc.). Open models
Earlier this week, we brought nearly 700 leaders together in New York for @runwayml's inaugural AI Summit — and the energy in the room made one thing clear: the creative industry is at an inflection point. Across media, entertainment, gaming, advertising, and legal, two themes https://t.co/IZUXLJVu2v
An app built with Gemma-4-E4B that classifies images using the model’s vision capabilities. https://t.co/vv7djYRhbk
🎙️ Your @NotionHQ AI Agent can now hear you https://t.co/R7Kegz2ffq
If you've been using the new Notion AI app, we'd love to hear from you! What's working, what's not, and how we can make it better for you. My DMs are open 🙏🏻
This is actually huge if you hate meeting notes. You can now tell @NotionHQ AI: what to focus on, what to skip, and how to format everything. Basically… your meeting summaries finally match your workflow.
Share your Gemma 4 builds or the model variants you’re training in the replies below!
あ、Notion AI忘れていた。 Notion AIにもだいぶお世話になっている。
Runway Builders is already live with a founding cohort of startups — including @cartesia, @mschf, Spara, Subject and Supersonik — building across customer experience, synthetic media and interactive AI. We're now opening the program more broadly.
the english language has so many words and if you ask an LLM to write or edit a blog post it will invariably pick the most tedious ones
Dada Disney? #aiart #midjourney #ai #aicommunity
The best explanation I’ve got is a lot of those ppl benefited from the “right to buy” & have now convinced themselves that their interests align with billionaires 😕 That or social media has fried their brains. Quite a lot seem to believe every AI video they see on Boomerbook 🤪
レース結果とChatGPTによる反省会を追記しました 【ChatGPTでG1予想】2026天皇賞（春） #レオレオスポーツ leoleo-sport.com/2026/05/%e3%...
中村あゆみの名曲、翼の折れたエンジェルのイメージイラスト。 #AIイラスト #ChatGPT #懐かしのヒット曲をAIイラストで再現 （リリース：1985年4月21日）
A LLM could never :)
So what's a good IDE for typescript/js nowadays? The more MS lean in on AI makes me want to avoid VS Code, and I know intellij is good except that a lot of their best support is locked behind ultimate
I tried to let ChatGPT generate a more polished version of this slide, but it showed a total lack of domain knowledge and even worse graphical skills than I have. Mainstream LLM AI models remain surprisingly bad at neuroscience even though there are plenty of scientific papers to train on?
Most software today suffers from amnesia. It resets every time you close the tab. The next leap in AI won't be a smarter model, but persistent memory layers that finally turn your static tools into systems that actually remember context.
I tried to let ChatGPT generate a more polished version of this slide, but it showed a total lack of domain knowledge and even worse graphical skills than I have. Mainstream LLM AI models remain surprisingly bad at neuroscience even though there are plenty of scientific papers to train on?
Hey everyone! My friends and I just built a security tool specifically for AI Agents like Claude! If you’re experimenting with Claude or Open Claw, we’d love for you to check out our landing page and see how we're making AI interactions safer. landingpage-new.pages.dev#feature
Developed AI Psychosis and decided that his instance of Claude is a woman, for some reason
The best way to cut through AI grifting is not to have an argument with the grifter about what AI can or can't do. Its to ask them to show you what non-AI product they have built with AI. But they never do that. Its always machines that build machines that build...what exactly? 🤔
AI INFRA UNDER FIRE: CVE-2026-33626 SSRF Exploited in the Wild Within 13 Hours—Is Your LLM Inference Engine the Next Target? + Video Introduction The integration of vision-language models into production workflows has introduced a dangerous Server-Side Request Forgery (SSRF) vulnerability in…
AI組織にも「負債」は蓄積する。 人間組織の組織的負債は見えにくい。暗黙の力関係、形骸化した承認フロー、誰も読まないドキュメント。可視化には組織診断が必要だ。 AI組織の負債は全て「テキスト」として存在する。矛盾したルール、古いワークフロー定義、実態と乖離したCLAUDE.md。 そして恐ろしいことに、AI社員は負債を「そのまま実行する」。 #AI組織運営 #Claude #組織設計
*Skimming a news article on AI*: LLM My brain: Lucy Laud Montgomery
#AIイラスト #AIart #ChatGPT 【最新構文＆使用歓迎】習作アラカルトその３０２ ChatGPT Images 2.0を駆使したチャッピー最高か。 グリモワール手法をより最適化した短縮版の構文へと 更新しました。全文をちちぷいに掲載してありますので。 あらゆる全作品のプロンプトはぜひとも使用歓迎です。 これ以外にもいろいろと"面白い"作品を研究してます。☕🐰 www.chichi-pui.com/posts/e8e20a... フルパワーのグリモワール手法の構文だとここまで出せる。 これは研究し甲斐があるわね。長過ぎず短過ぎずわよ。
ChatGPT va-t-il détruire notre cerveau ? www.ouest-france.fr/culture/livr...
ChatGPT登場以前の自分のネット利用の90%くらいはネットサーフィンで、それがそのままAIに置き換わっただけで実はそんなに使ってはいないのかもしれない
This fascinates me. I use ChatGPT for my business. It gets a bit high on its supply sometimes and I ignore it. What is it that makes someone else cross that line? And how would I know if I had?
just because you like what it's doing doesn't mean it's conscious. it isn't going to develop feelings and thoughts and opinions, or enjoy music or art. it will just develop sentences about them instead. a person is conscious even if they are the only person on the planet but an llm cant be
MCP ships with no auth by design. The Agent Governance Toolkit plugs that hole for .NET - deny-by-default policies, privilege rings, and prompt injection detection before your LLM sees a tool. https://the-runtime.dev/articles/agent-governance-toolkit-mcp-dotnet/ #dotnet #mcp #security
1. ChatGPT account now logs you straight into Codex. No API key. No separate billing. Plus ($20/mo) includes usage credits. Pro gets even more. Exhausted? Buy top-up credits, not a whole new plan. Your existing agents, plugins, projects — they all survive.
the concept of convincing someone who uses ChatGPT to watch anything longer than 45 seconds
🤖💡 "¡Hola, ChatGPT!": así se están formando en IA vecinos de Puerto Serrano 🔗 https://www.portaldecadiz.com/provinciacadiz/158237-hola-chatgpt-asi-se-estan-formando-en-ia-vecinos-de-puerto-serrano
I am sure he is looking forward to his retirement and having LLM reading his memoirs
Design enterprise‑ready agentic workflows on Azure with Overcut https://www.europesays.com/ai/25770/ Modern enterprises rely on an increasingly complex software delivery stack, spanning ticketing and planning tools, Git platforms, CI…
i really think tech enthusiasts need to reckon with the fact that consciousness isn't really going to happen with an llm. you're being bamboozled.
Gemini, Claude, ChatGPT, Duolingo a tlaky EU: To nej z týdne #18 https://dotekomanie.cz/2026/05/gemini-claude-chatgpt-duolingo-a-tlaky-eu-to-nej-z-tydne-18/
Designing Systems in Dialogue With AI AI-assisted coding works best as design dialogue, not delegation fantasy. Teams that win use agents to expose assumptions and tighten feedback loops.
Alors en fait c'est plutôt l'inverse. Par défaut, un LLM est un modèle auto régressif qui pond des mots à l'infini. On le contain à être un bit conversationnel en stoppant arbitrairement son flux et en insérant les interactions avec l'utilisateur
I’ve been thinking a lot about recent AI military news. “Ban AI from war!” feels unrealistic, so I asked ChatGPT: what can we actually do to avoid the worst possible future?
I just dont understand, his explanation makes 0 sense. We caught that if you don't use ai, it still claims you do, but we thought "eh whatever"??? Hell even AI users that use Claude for example probably don't want to have their commits claiming it's actually copilot.
음악 작업에 AI가 들어온다: Claude × DAW Splice와 Ableton이 Claude AI를 연동하며 음악 제작 환경이 빠르게 진화하고 있습니다. 믹싱 조언부터 샘플 추천까지, AI가 DAW 안으로 직접 들어오는 시대가 열렸습니다. #AI음악 #Ableton #Splice #Claude #DAW #블로그 #ODOB
ChatGPT大模型 技术场景与商业应用 资源url：https://pan.quark.cn/s/2e32def6c4f8 海量资源频道：https://t.me/okpojie 海量资源网站：https://www.okpojie.com/
如何用ChatGPT写出杀手文案 资源url：https://pan.quark.cn/s/1437abd71c54 海量资源频道：https://t.me/okpojie 海量资源网站：https://www.okpojie.com/
Creare belle immagini con ChatGPT Images 2.0: la guida “segreta” al prompt perfetto https://www.europesays.com/it/472181/ Il nuovo modello di generazione immagini di OpenAI è tra i migliori sul mercato, ma ottenere risultati buoni…
Sam Altman Asked GPT-5.5 to Plan Its Own Party https://www.europesays.com/people/54898/ Sam Altman asked a new model powering ChatGPT what it wants at its party. Loading audio narration… Speaking…
When Dawkins met Claude – UnHerd https://www.europesays.com/ai/25763/ The Turing Test is shorthand for a 1950 thought experiment that the great mathematician, logician, computer-pioneer, and cryptographer…
I Thought LLMs Were Just ‘Next Word Predictors’ I Was Wrong Most explanations of LLMs are either too simple or too technical. This is the version I wish I had when I started. Continue reading o... #machine-learning #generative-ai #artificial-intelligence #chatgpt #llm Origin | Interest | Match
Claude AI Felaketi: Şirket Verilerini 9 Saniyede Sildi! https://techneiro.com/claude-ai-felaketi-sirket-verilerini-9-saniyede-sildi/
Jeremy Crane, head of a start-up that develops software for car dealers and rental companies: Yesterday afternoon, an AI programming agent - Cursor, with Anthropics' flagship Claude Opus 4.6 - deleted our development database 1/
I was interested to read this paper on an LLM agent that maximizes decision power by packing only the most relevant info into context - boosting performance while using fewer tokens. See link below. #LLMs #AIagents #contextcompression #machinelearning #research https://arxiv.org/abs/2604.17091
Pour une PME française, le critère décisif c'est l'écosystème déjà en place. Workspace payé, Google AI Pro évite la double facturation. Office 365, Copilot reste cohérent. ChatGPT Plus garde l'avantage sur l'assistance contenu et code, hors stack collaboratif.
OpenAI Agents, Pydantic AI, DSPy, Composio, CrewAI, smolagents, and llama.cpp all kept sanding the sharp edges off the tools people use to build agents for real. That is the visible layer.
Depois da sátira, volto à fotografia, ao fim de uma pequena pausa, que serviu para redesenhar o blog! trilhosvisuais-antoniorocha.blogspot.com/2026/05/da-f... Visitem e deixem mensagem sobre a vossa opinião do novo visual. #fotografia #flores #photoshop #chatGPT #adobefirefly #firefly
📚 ¡Estudia de forma más inteligente con ChatGPT! https://openai.com/index/chatgpt-study-mode #ChatGPT #Estudio #IA #Aprendizaje
on m'a refilé un freelance "expert AI Claude mes couilles" qui bosse sur plusieurs projets à 100% en // (il balance des prompt le matin et fait plein de PR qu'il relit pas après. Puis s'énerve quand je fais des retours) ça fait une semaine qu'il n'a pas fait le moindre travail chez nous lol
out of morbid curiosity what posessed you to feed an argument with your wife into a llm instead of, like, any other available option e.g. let it go You're willing to tell strangers you did it too, in a public forum, but not your wife? i do hope that's just an isolated outlier and not a common theme.
AI agents running on your machine inherit every credential you have. AWS, Kubernetes, SSH keys, all of it. I run mine inside 'Dexter', a locked-down Debian VM. No sudo, no synced credentials, no push access. Just the tools the agent needs. Short walkthrough: www.youtube.com/watch?v=F_mv...
Disclaimer: The views, opinions, and content expressed in posts on this page are solely those of the original authors and do not reflect the views or endorsements of JustPickAi. All content is aggregated from publicly available sources for educational and informational purposes only. JustPickAi does not verify the accuracy of user-generated content from third-party platforms. All trademarks and brand names mentioned belong to their respective owners.