Wiredframe Radar logo

Category intelligence

Social Media Briefing — May 20, 2026

610 current items analyzed and ranked.

Executive synthesis

Social Media Summary

Google I/O dominated the day with two major model announcements: Gemini 3.5 Flash (frontier coding/agentic performance at flash-tier pricing) and Gemini Omni (multimodal generation starting with video). Google also unveiled a unified AI Search experience rolling out to 1.5B users.

  • Andrej Karpathy announced joining Anthropic, one of the biggest talent moves in AI history, signaling Anthropic's growing gravitational pull on top researchers
  • Sam Altman revealed OpenAI is offering discounted tokens for 1-3 year capacity commitments, predicting persistent compute constraints as models improve
  • Ethan Mollick shared peer-reviewed PNAS research showing classic human persuasion techniques work on LLMs in 'parahuman' ways, raising alignment concerns
  • François Chollet argued most real-world tasks are non-Markovian, challenging current agentic AI architectures that rely on present-state observation
  • NVIDIA released Nemotron-Labs-Diffusion, a novel family of diffusion language models that generate multiple tokens in parallel, offering an alternative to autoregressive generation

Key Themes

Karpathy Joins Anthropic · 13Google I/O & Gemini 3.5 Launch · 18Google I/O: Gemini 3.5 Flash & Omni · 16Gemini Omni Video Generation · 12OpenAI Capacity Commitments · 4New Google Model Releases (Gemini 3.5 Flash & Gemini Omni) · 5Google IO 2026 Coverage · 4AI Safety & Vulnerabilities · 1Google I/O & Gemini Omni · 2Open Science & Brain Drain · 6

Primary evidence

Top Ranked Signals

Social Twitter May 19

Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be e...

By @karpathy

97 score
AI Analysis

Andrej Karpathy announces he has joined Anthropic, citing excitement about the next few years at the LLM frontier, while maintaining his passion for education.

Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.
talent_movementanthropicindustry_dynamics
Social Twitter May 19

Introducing Gemini 3.5: our newest family of models combining frontier intelligence with real-world ...

By @GoogleDeepMind

40 score
AI Analysis

Google DeepMind introduces Gemini 3.5 model family combining frontier intelligence with real-world action, starting with 3.5 Flash as strongest model for agents and coding

Introducing Gemini 3.5: our newest family of models combining frontier intelligence with real-world action. The first release is 3.5 Flash, our strongest model yet for agents and coding 🧵
gemini_3.5model_releaseagentic_aicodinggoogle_io
Social Twitter May 19

customers are increasingly asking us for certainty on capacity. as models get better, we expect that...

By @sama

90 score
AI Analysis

Sam Altman announces OpenAI offering discounted tokens for 1-3 year capacity commitments, predicting capacity constraints will persist as models improve.

customers are increasingly asking us for certainty on capacity. as models get better, we expect that the world will be capacity-constrained for some time. we are offering discounted tokens for 1-3 year commits. (it also helps us plan, so hopefully a big win-win.)
openai_businesscapacity_constraintsenterprise_aipricing_strategycompute_scaling
Social Twitter May 19

We’re dropping Gemini Omni: our first step towards a model that can create anything from anything - ...

By @GoogleDeepMind

90 score
AI Analysis

Google DeepMind announces Gemini Omni: first step toward a model that creates anything from anything, starting with video. Combines Gemini intelligence with generative media systems.

We’re dropping Gemini Omni: our first step towards a model that can create anything from anything - starting with video. It combines Gemini’s intelligence with our generative media systems - representing a leap forward in world understanding, multimodality, and editing 🧵
gemini_omnivideo_generationmodel_releasemultimodal_aigoogle_io
Social Twitter May 19

Gemini Omni is a major leap in world understanding & multimodal editing! It can take photos, vid...

By @demishassabis

88 score
AI Analysis

Demis Hassabis announces Gemini Omni as a 'major leap in world understanding & multimodal editing' - can take photos/video/audio and build new scenes, with iterative editing capability and path to any-input any-output.

Gemini Omni is a major leap in world understanding & multimodal editing! It can take photos, video & audio and build entirely new scenes. Over time it’ll be able to handle any input & any output - starting w/ video You can even give it your own videos & iterate on your ideas: t.co/VrHPJKRJXH
google_gemini_omnimultimodal_aivideo_generation_progressmodel_releasesgoogle_io
Social Twitter May 19

Today, we launched a brand-new intelligent Search box. Here's what that means: An upgrade to the ...

By @GoogleAI

88 score
AI Analysis

Google AI announces new intelligent Search box powered by Gemini 3.5 models, combining AI Overviews and AI Mode into one unified AI Search experience with multimodal capabilities, live worldwide

Today, we launched a brand-new intelligent Search box. Here's what that means: An upgrade to the Search experience with our most advanced Gemini 3.5 models, bringing with them our latest agentic capabilities You can ask across modalities (text, images, files, and videos) and Search can reason across them all We're combining AI Overviews and AI Mode into one, seamless AI Search experience. So you can ask follow-up questions, build context, and received even more tailored and personalized res
google_iogemini_3.5ai_searchproduct_launch
Social Twitter May 19

1/ Today at #GoogleIO, we’re releasing Gemini 3.5, our latest family of models combining frontier in...

By @JeffDean

85 score
AI Analysis

Jeff Dean's main Google I/O announcement: Gemini 3.5 Flash released, strongest model for coding/agents, outscores 3.1 Pro on Terminal-Bench and MCP Atlas, 4x faster than frontier models, 12x faster in Antigravity, designed for collaborative sub-agents.

1/ Today at #GoogleIO, we’re releasing Gemini 3.5, our latest family of models combining frontier intelligence with action. We’re starting by releasing 3.5 Flash, which is built to help you execute complex, long-horizon agentic workflows. Gemini 3.5 Flash is our strongest model for coding and agent t.co/m62cBJhIjJ outscores 3.1 Pro on agentic and coding benchmarks like Terminal-Bench and MCP Atlas, while running 4x faster than other frontier models. Used in Google Antigravity, 3.5 Fla
google_geminimodel_releasescoding_agentsagentic_aigoogle_iobenchmarks
Social Bluesky May 19

My notes on Gemini 3.5 Flash - 3x the price of Gemini 3 Flash but Google are planning to use it for ...

By @simonwillison.net

83 score
AI Analysis

Simon Willison shares detailed notes on Gemini 3.5 Flash, noting it's 3x the price of Gemini 3 Flash and that Google plans to use it internally for many products.

My notes on Gemini 3.5 Flash - 3x the price of Gemini 3 Flash but Google are planning to use it for many of their own products simonwillison.net/2026/May/19/...
new_model_releasegoogle_geminimodel_pricingproduct_strategy
Social Twitter May 19

we are offering discounted tokens and certainty on capacity availability in exchange for 1-3 year co...

By @gdb

40 score
AI Analysis

OpenAI's Greg Brockman announces discounted tokens and capacity guarantees for 1-3 year commitments, predicting the world will be capacity-constrained as models improve.

we are offering discounted tokens and certainty on capacity availability in exchange for 1-3 year commits. we expect that the world will feel increasingly capacity constrained for the next while, as models continue to get much more useful.
openai_businesscapacity_constraintsenterprise_aipricing_strategy
Social Twitter May 19

Introducing OpenAI Guaranteed Capacity: a new offering that enables customers to guarantee long-term...

By @OpenAI

82 score
AI Analysis

OpenAI introduces Guaranteed Capacity: a new enterprise offering allowing customers to guarantee long-term access to OpenAI compute for critical workloads

Introducing OpenAI Guaranteed Capacity: a new offering that enables customers to guarantee long-term access to OpenAI compute. We’ve made long-term investments in infrastructure, partnerships, and capacity planning to help customers scale reliably. Now, Guaranteed Capacity helps customers plan ahead for critical workloads in a compute-constrained world. t.co/TN4OkZr2Uo
openai_enterprisecompute_infrastructurebusiness_strategy
Social Twitter May 19

One of the loudest applauses in the entire Google keynote: Nishtha put on the new Gentle Monster + ...

By @alliekmiller

82 score
AI Analysis

Allie K Miller reports on Google IO keynote highlight: Gentle Monster + Gemini glasses demo where user summons Gemini via tap, gives one compound prompt to take photo and add cartoon blimp saying 'Google IO 2026', and edited photo appears on watch within seconds via 'nano banana'. Notes loud applause and AI coming everywhere via voice interaction.

One of the loudest applauses in the entire Google keynote: Nishtha put on the new Gentle Monster + Gemini glasses, tapped the side to summon Gemini, and ALL in one prompt said “take a photo and put a cartoon blimp in the sky that says Google IO 2026” and within seconds, the preview of the edited photo from nano banana appeared on her watch. I want to spend less time on screens. AI really is coming everywhere. And so much is driven by voice AI as the interaction mode. #Google
Google IO 2026Gemini glassesmultimodal AIAI wearablesvoice AIVision-Banana