Wiredframe Radar logo

Category intelligence

Social Media Briefing — May 23, 2026

566 current items analyzed and ranked.

Executive synthesis

Social Media Summary

OpenAI's Codex Thursday and a strategic pivot dominated discussions, with Greg Brockman declaring 'the model alone is no longer the product' as Codex gained remote Mac control from phones. The framing of harnesses and integrations as the new battleground resonated widely.

Key Themes

Codex Thursday & 'Model is no longer the product' · 11AI policy & immigration · 1Project Glasswing & AI Cybersecurity · 2NVIDIA Research Releases · 8Open-Source Developer Security & Supply Chain · 3Benchmark Coverage & Evaluation (AI2 ArtifactLinker) · 4Multimodal Video AI · 1AI Agents & Coding Tools · 10Gemini Omni multimodal video · 2AI for Scientific Discovery · 1

Primary evidence

Top Ranked Signals

Social Twitter May 22

Last month we launched Project Glasswing, our collaborative AI cybersecurity initiative. Since then,...

By @AnthropicAI

85 score
AI Analysis

Anthropic announces Project Glasswing has found 10,000+ high/critical-severity vulnerabilities in essential software since launch.

Last month we launched Project Glasswing, our collaborative AI cybersecurity initiative. Since then, we and our partners have found more than ten thousand high- or critical-severity vulnerabilities in essential software.
AI securityProject GlasswingAnthropicvulnerabilities
Social Twitter May 22

The new White House policy requiring green card applicants to apply from outside the US is a caprici...

By @AndrewYNg

80 score
AI Analysis

Andrew Ng criticizes White House green card policy as harmful to AI competitiveness.

The new White House policy requiring green card applicants to apply from outside the US is a capricious attack on legal immigration. It will hurt families, leave us with fewer doctors, teachers and scientists, and hurt American competitiveness in AI.
AI policyimmigrationtalent
Social Twitter May 22

the model alone is no longer the product

By @gdb

80 score
AI Analysis

Greg Brockman declares 'the model alone is no longer the product' - signaling shift to integrated AI systems/harnesses.

the model alone is no longer the product
AI productsmodel strategyOpenAI
Social Twitter May 22

I think people don't realize why Gemini Omni is different than other video AIs. It is fully multimod...

By @emollick

78 score
AI Analysis

Ethan Mollick demos Gemini Omni's native multimodal video editing - editing 1896 train film to add bullet train, LEGO, time travelers, muppets.

I think people don't realize why Gemini Omni is different than other video AIs. It is fully multimodal, so it can edit video natively, too I took the famous "train " movie from 1896 & made it a bullet train, LEGO, added a time traveler, a centipede, muppets... (see reflections?) t.co/WxSO5h9x1Z
Gemini Omnimultimodal videoAI demos
Social Twitter May 22

Today we're open-sourcing Bumblebee, a read-only scanner for macOS and Linux. It checks developer m...

By @perplexity_ai

78 score
AI Analysis

Perplexity open-sources Bumblebee, a read-only scanner for macOS/Linux developer machines that checks for risky packages, extensions, and AI tool configs.

Today we're open-sourcing Bumblebee, a read-only scanner for macOS and Linux. It checks developer machines for risky packages, extensions, and AI tool configs. Connected to Computer, it can trigger deeper scans whenever a new supply-chain risk emerges. t.co/FOaWnF1yQy t.co/wXauD4wDOT
securityopen sourcedeveloper toolssupply chain
Social Twitter May 22

babyagi has ~200 citations, but 0 papers... i just published my first paper on arXiv 😆 "The Log is ...

By @yoheinakajima

78 score
AI Analysis

Yohei Nakajima publishes first arXiv paper 'The Log is the Agent: Event-Sourced Reactive Graphs for Auditable, Forkable Agentic Systems', advocating agents that coordinate via persistent replayable state.

babyagi has ~200 citations, but 0 papers... i just published my first paper on arXiv 😆 "The Log is the Agent: Event-Sourced Reactive Graphs for Auditable, Forkable Agentic Systems" t.co/c7mbRggdCh the case for agents that coordinate through persistent replayable state — no conversation loops, no workflows, no A2A — with auditability, forking, and causal lineage built in. check it out and let me know what you think!
agentsevent sourcingBabyAGIpapers
Social Twitter May 22

Highlights from today’s Codex Thursday launches: 1️⃣ Codex can now securely use apps on your Mac fr...

By @OpenAI

75 score
AI Analysis

Following yesterday's Social teaser from Sam Altman, OpenAI Codex Thursday launch: Codex can securely use apps on your Mac from your phone even when locked, plus other features.

Highlights from today’s Codex Thursday launches: 1️⃣ Codex can now securely use apps on your Mac from your phone, even when your Mac is locked and the screen is off. t.co/JUOss3M2Va t.co/CAEUe1aswm
Codexagentsremote controlOpenAI
Social Twitter May 22

Long video generation is a systems problem. Introducing LongLive-2.0 from NVIDIA Research: an end-t...

By @NVIDIAAI

75 score
AI Analysis

NVIDIA Research introduces LongLive-2.0, an end-to-end NVFP4 training and inference system aligning low-precision training and W4A4 inference for long video generation.

Long video generation is a systems problem. Introducing LongLive-2.0 from NVIDIA Research: an end-to-end NVFP4 training and inference system for long video generation. Low-precision deployment often relies on post-training quantization, creating a gap between how models are trained and how they run. LongLive-2.0 aligns NVFP4-aware training, distillation, and W4A4 inference, maintaining strong benchmark quality while improving speed and memory efficiency.
video generationlow-precision trainingNVFP4NVIDIA research
Social Twitter May 22

Most models are only evaluated on a fraction of the benchmarks out there. ArtifactLinker, our new s...

By @allen_ai

75 score
AI Analysis

AI2 announces ArtifactLinker, a system that predicts which models would set new SOTA on HuggingFace benchmarks and verifies via running evals.

Most models are only evaluated on a fraction of the benchmarks out there. ArtifactLinker, our new system, predicts which ones would set a new state-of-the-art on benchmarks hosted on @HuggingFace, then runs the evaluation to verify. 🧵 t.co/DXk3FbndMF
benchmarkingevaluationAI2 research
Social Bluesky May 22

I think people don't realize why Gemini Omni is different than other video AIs. It is fully multimod...

By @emollick.bsky.social

75 score
AI Analysis

Mollick demonstrates Gemini Omni's native multimodal video editing by transforming the 1896 'train' film with various creative elements.

I think people don't realize why Gemini Omni is different than other video AIs. It is fully multimodal, so it can edit video natively, too I took the famous "train " movie from 1896 & made it a bullet train, LEGO, added a time traveler, a centipede, muppets... (note the reflections and shadows?)
multimodal-AIvideo-generationGemini
Social Twitter May 22

Say hello to open source deep research for your favorite agent harness. Our AI-Q agent skill packag...

By @NVIDIAAI

72 score
AI Analysis

NVIDIA introduces AI-Q, an open-source deep research agent skill that packages a research pipeline as a portable skill droppable into agent harnesses (e.g., Codex).

Say hello to open source deep research for your favorite agent harness. Our AI-Q agent skill packages the work of building a research pipeline into a portable skill. Drop it into your harness, and the agent delegates a research task to a local or hosted AI-Q server and gets back a detailed report with citations. See it in Codex below 👇
AI agentsdeep researchopen sourceNVIDIA
Social Twitter May 22

I am thrilled that my article in @americanacad Daedalus special issue on AI & Science: What Is the ...

By @AnimaAnandkumar

70 score
AI Analysis

Anima Anandkumar shares Daedalus article on AI for Scientific Discovery, arguing AI's transformative role lies in granular physics simulation (Neural Operators) to reduce experimentation, not just abstract textual reasoning.

I am thrilled that my article in @americanacad Daedalus special issue on AI & Science: What Is the Future of Discovery? edited by James Manyika. t.co/vvur95HXGI I talk about : How Do We Build AI to Push the Frontiers of Scientific Discovery? Scientific progress is limited not by a lack of new ideas but by the time and cost involved in physical experimentation. Scientific discovery is a needle in the haystack problem: it does not help if AI gives you a vastly bigger haystack. Without k
AI for ScienceNeural Operatorsscientific discovery