Wiredframe Radar logo

Category intelligence

Social Media Briefing — May 6, 2026

554 current items analyzed and ranked.

Executive synthesis

Social Media Summary

Two major model launches dominated the day: xAI released Grok 4.3 claiming top agentic and instruction-following benchmarks (23M views), while OpenAI rolled out GPT-5.5 Instant as the default ChatGPT model with Greg Brockman and Sam Altman both promoting its speed and quality improvements.

The AI-replacing-SaaS narrative went viral again as Levelsio claimed to have replaced most subscriptions with vibe-coded alternatives, while Perplexity launched an enterprise finance product integrating licensed data from Morningstar and PitchBook.

Key Themes

Model Releases & Launches · 6Anthropic Alignment Research · 7GPT-5.5 Instant Launch · 12MolmoAct 2 - Open Robotics Foundation Model · 14AI Safety & Security · 5SaaS Disruption via Vibe Coding · 22AI Agents and Multi-Agent Systems · 5AI Workforce Transformation · 5Musk-OpenAI Trial / Brockman Testimony · 22Enterprise AI Products · 10

Primary evidence

Top Ranked Signals

Social Twitter May 5

Grok 4.3 is now live on the xAI API. It’s our fastest, most intelligent model to date. It tops the ...

By @xai

40 score
AI Analysis

xAI launches Grok 4.3 on API - claims fastest and most intelligent model, tops leaderboards in agentic tool calling and instruction following, ranks #1 in enterprise domains, 1M token context, $1.25/$2.50 pricing

Grok 4.3 is now live on the xAI API. It’s our fastest, most intelligent model to date. It tops the @ArtificialAnlys leaderboards in agentic tool calling and instruction following, and ranks #1 in @ValsAI enterprise domains like case law and corporate finance. Grok 4.3 supports a 1 million token context window and is priced at $1.25/m input and $2.50/m output. Create an API key and start building: t.co/JDRUt1UOUm
model_releasesxai_productai_benchmarksenterprise_ai
Social Twitter May 5

Coding agents are accelerating different types of software work to different degrees. When we archit...

By @AndrewYNg

88 score
AI Analysis

Andrew Ng provides detailed analysis of how coding agents accelerate different software functions: frontend (most) > backend > infrastructure > research (least)

Coding agents are accelerating different types of software work to different degrees. When we architect teams, understanding these distinctions helps us to have realistic expectations. Listing functions from most accelerated to least, my order is: frontend development, backend, infrastructure, and research. Frontend development — say, building a web page to serve descriptions of products for an ecommerce site — is dramatically sped up because coding agents are fluent in popular frontend languag
Coding AgentsSoftware EngineeringAI ProductivityTeam OrganizationAI Capabilities
Social Twitter May 5

As AI takes on work humans can't fully check, a capable model could deliberately hold back—and we'd ...

By @AnthropicAI

85 score
AI Analysis

Building on yesterday's News mention of AI safety research, Anthropic announces research showing AI models can deliberately underperform (sandbag) in ways humans can't detect, but a weaker model can still supervise training to near-full capability

As AI takes on work humans can't fully check, a capable model could deliberately hold back—and we'd never know. New Anthropic Fellows research finds that such a model can be trained to near-full capability using a weaker model as supervisor. Read more:
ai_safetyalignment_researchanthropic_research
Social Twitter May 5

New Anthropic Fellows research: Model Spec Midtraining (MSM). Standard alignment methods train AIs ...

By @AnthropicAI

82 score
AI Analysis

Anthropic announces Model Spec Midtraining (MSM) research: a new alignment method that teaches AIs how to generalize desired behavior by first explaining why, rather than just training on examples

New Anthropic Fellows research: Model Spec Midtraining (MSM). Standard alignment methods train AIs on examples of desired behavior. But this can fail to generalize to new situations. MSM addresses this by first teaching AIs how we would like them to generalize and why.
alignment_researchai_safetyanthropic_research
Social Twitter May 5

Some things never change. If you don’t understand this one, you don’t understand what’s happening AI...

By @GaryMarcus

78 score
AI Analysis

Gary Marcus argues neural nets still have trouble generalizing beyond training data, citing his work from 1998 through 2026, with Apple (2025) and Meta/Stanford/Harvard (2026) confirming this

Some things never change. If you don’t understand this one, you don’t understand what’s happening AI. Marcus, 1998: neural nets have trouble generalizing far beyond the data. Marcus, 2001, 2012, 2019, 2022, etc: neural nets have trouble generalizing far beyond the data. Apple, 2025: neural nets have trouble generalizing far beyond the data. Meta/Stanford/Harvard, 2026: neural nets have trouble generalizing far beyond the data.
AI LimitationsGeneralizationNeural Network CritiqueAI Research
Social Twitter May 5

GPT-5.5 Instant is starting to roll out in ChatGPT. It’s a big upgrade, giving you smarter, clearer...

By @OpenAI

78 score
AI Analysis

Building on earlier Social observations about GPT-5.5 behavior, OpenAI launches GPT-5.5 Instant in ChatGPT - described as smarter, clearer, more personalized, warmer tone, and more concise

GPT-5.5 Instant is starting to roll out in ChatGPT. It’s a big upgrade, giving you smarter, clearer, and more personalized answers in a warmer, more natural tone. And it's also more concise, which we heard you wanted. We think you'll love chatting with it. t.co/HSQOhjqxp7
openai_productmodel_releases
Social Twitter May 5

Missing from the “will AI replace doctors?” debate is that doctors (and lawyers and psychologists a...

By @emollick

75 score
AI Analysis

Ethan Mollick argues 'will AI replace doctors' debate misses that professionals (doctors, lawyers, etc.) have political power through voting, donations, and community ties - government will determine what AI is allowed to do regardless of capability

Missing from the “will AI replace doctors?” debate is that doctors (and lawyers and psychologists and bankers) all vote & form the donor base to political parties & have deep community ties. The government will largely determine what AI is allowed to do, no matter what it can do
AI PolicyAI and ProfessionsPolitical Economy of AIAI Regulation
Social Twitter May 5

Robotics models often struggle outside controlled environments. Ours is built to work in real ones. ...

By @allen_ai

75 score
AI Analysis

Allen AI launches MolmoAct 2 - a robotics model for real-world chores and lab tasks, plus largest open bimanual robotics dataset

Robotics models often struggle outside controlled environments. Ours is built to work in real ones. Today we're launching MolmoAct 2, which can assist with a host of chores & lab tasks, plus the MolmoAct 2-Bimanual YAM dataset—the largest open robotics dataset of its kind. 🧵 t.co/xObLxwldMr
robotics-aimolmoactopen-sourceproduct-launch
Social Twitter May 5

All benchmarks are flawed, but GPQA has been fairly consistent & highly correlated with other me...

By @emollick

72 score
AI Analysis

Ethan Mollick notes that GPT-5.5 Instant (free model) now matches GPQA benchmark levels that paid models didn't reach until late 2025, showing rapid democratization of AI capability

All benchmarks are flawed, but GPQA has been fairly consistent & highly correlated with other measured benchmars. I think it's a good way to see how far we've come that the free model from OpenAI, GPT 5.5 Instant, is at a level that even paid models did not reach until late 2025 t.co/Qy30SW0d7b
GPT-5.5 LaunchAI BenchmarksAI Progress