Wiredframe Radar logo

Category intelligence

Social Media Briefing — April 7, 2026

466 current items analyzed and ranked.

Executive synthesis

Social Media Summary

Anthropic dominated the day with two blockbuster announcements: a multi-gigawatt TPU deal with Google and Broadcom for 2027, and run-rate revenue surging from $9B to $30B. Meanwhile, an Anthropic engineer publicly investigated why Claude Code MAX 20x users burn tokens unexpectedly, finding subagent spawning—not prompt caching—is the culprit.

  • A viral New Yorker exposé on Sam Altman drew massive attention (217K views), while Altman himself published a 13-page policy blueprint proposing robot taxes, a public wealth fund, and a 4-day workweek
  • Clément Delangue (HuggingFace) launched an initiative to crowdsource agent traces for open-source frontier agents, identifying data as the key bottleneck
  • François Chollet articulated a fundamental distinction between lossy curve-fitting and lossless symbolic program synthesis, arguing symbolic approaches will dominate for simple latent programs
  • Research showing GPT-4o, Gemini, and DeepSeek reproduce up to 90% of copyrighted books after fine-tuning was called a 'smoking gun' for ongoing copyright lawsuits
  • Andrej Karpathy shared practical wisdom on knowledge work: LLMs let you skip writing but never reading and thinking (80K views)
  • Ethan Mollick offered a sobering take that no large firm likely saw major GenAI work impacts throughout 2025, framing 2026 as the true inflection point

Key Themes

Anthropic Business & Infrastructure · 4Claude Code Token Usage Investigation · 22Sam Altman / New Yorker Investigation · 15Open-Source Agent Data Initiative · 3Symbolic Learning vs Curve-Fitting · 1AI Policy & Superintelligence Preparation · 4Claude Code Issues & Anthropic Response · 4GenAI Enterprise Adoption Timeline · 1Gemma 4 Safety Guardrails Criticism · 4LLM-Assisted Knowledge Management · 1

Primary evidence

Top Ranked Signals

Social Twitter Apr 6

We've signed an agreement with Google and Broadcom for multiple gigawatts of next-generation TPU cap...

By @AnthropicAI

97 score
AI Analysis

Anthropic announces a major deal with Google and Broadcom for multiple gigawatts of next-generation TPU capacity starting in 2027 to train and serve frontier Claude models.

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.
compute_infrastructureanthropic_businessgoogle_partnershipai_scaling
Social Twitter Apr 6

Our run-rate revenue has surpassed $30 billion, up from $9 billion at the end of 2025, as demand for...

By @AnthropicAI

95 score
AI Analysis

Following yesterday's Reddit reports of Anthropic surpassing OpenAI in ARR at $25B, Anthropic announces run-rate revenue has surpassed $30 billion, up from $9 billion at end of 2025, citing accelerating demand for Claude. Linked to compute partnership announcement.

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: t.co/XgSjL0And7
anthropic_businessai_revenueai_industry_growthcompute_infrastructure
Social Twitter Apr 6

I want to do a few more of these calls. If your MAX 20x plan ran out of tokens unexpectedly early...

By @trq212

88 score
AI Analysis

Anthropic's @trq212 (likely a Claude Code/product lead) publicly asks MAX 20x plan users who ran out of tokens unexpectedly early to do live screenshare debugging sessions. Mentions wanting to improve /usage to give more info. Extremely high engagement (248K views, 1.4K likes, 296 replies).

I want to do a few more of these calls. If your MAX 20x plan ran out of tokens unexpectedly early and you're willing to screenshare and run some prompts through Claude Code please comment. Trying to figure out how we can improve /usage to give more info.
Claude Code token usageAnthropic product issuesAI pricing and usage limitsDeveloper experience
Social Twitter Apr 6

We keep saying we want open-source frontier agents. Fine. Then let’s build the dataset. @badlogicg...

By @ClementDelangue

78 score
AI Analysis

Clement Delangue announces initiative to crowdsource agent traces for open-source frontier agent development. Shares his own traces from Hermes, opencode, and Claude. Argues the biggest bottleneck for open-source agents is data, and builders generate this data daily.

We keep saying we want open-source frontier agents. Fine. Then let’s build the dataset. @badlogicgames, creator of Pi, just shared some of his agent traces used to build Pi on @huggingface. I’m now sharing some of mine too, exporting them from @hermes, @opencode, and Claude via @tracesdotcom, and I’ll keep going. Why this matters: one of the biggest bottlenecks for open-source agent models is the data. And all of us are generating that data every day through our conversations with agents. If
open-source-AIagent-tracescrowdsourced-dataAI-agentsHuggingFacedata-bottleneck
Social Twitter Apr 6

Sam Altman just published a 13-page policy blueprint called "Industrial Policy for the Intelligence ...

By @TheRundownAI

78 score
AI Analysis

Sam Altman published a 13-page policy blueprint 'Industrial Policy for the Intelligence Age' proposing: public wealth fund, robot taxes, 4-day workweek, right to AI access, containment playbooks for autonomous AI, and auto-triggering safety nets for AI displacement.

Sam Altman just published a 13-page policy blueprint called "Industrial Policy for the Intelligence Age: Ideas to keep people first." The premise: AI superintelligence is so close that America needs a new social contract. The six proposals: > A Public Wealth Fund where every American gets a stake in AI growth > Robot taxes to replace payroll revenue AI is about to hollow out > A 4-day workweek at full pay, funded by AI efficiency gains > Right to AI, making access as foundational as liter
ai_policyopenaisuperintelligencelabor_displacementai_safetyubi_wealth_redistribution
Social Twitter Apr 6

Tried Gemma 4 ran locally on my iPhone today I thought it'd be useful in case the apocalypse happen...

By @levelsio

78 score
AI Analysis

Levelsio tested Gemma 4 running locally on iPhone, found it couldn't even answer how to make a fire. Humorous critique of the model's safety guardrails being overly restrictive.

Tried Gemma 4 ran locally on my iPhone today I thought it'd be useful in case the apocalypse happens and I need to ask it for survival tips Like how to make a fire 🔥 I guess I'll freeze to death instead 🫠 t.co/lFHKm5mcnd
gemma_4on_device_aisafety_guardrailsmodel_evaluationgoogle_ai
Social Twitter Apr 6

With curve-fitting, you are recording a lossy approximation of the output of some generative program...

By @fchollet

75 score
AI Analysis

François Chollet contrasts curve-fitting (lossy approximation) with symbolic learning (losslessly reverse-engineering the generative program). Argues symbolic learning will outperform by orders of magnitude where latent programs are simple.

With curve-fitting, you are recording a lossy approximation of the output of some generative program. With symbolic learning, you are losslessly reverse-engineering the source code of the generative program. Symbolic learning won't be the best fit for all problems, but for the ones where the latent program is reasonably simple, it will outperform by many orders of magnitude.
symbolic-learningML-theorycurve-fitting-vs-symbolicAI-research-directions
Social Twitter Apr 6

✅ Done 💳 Made an auto-dispute response system for Interior AI to see how easy it'd be It syncs old...

By @levelsio

75 score
AI Analysis

Levelsio shares completed auto-dispute-response system for Stripe built via vibe coding. Detailed prompt/skill for building a webhook-based system that auto-collects evidence, generates PDFs with customer usage data and product images, submits to Stripe, includes dashboard and Telegram notifications.

✅ Done 💳 Made an auto-dispute response system for Interior AI to see how easy it'd be It syncs old disputes but also catches new disputes via Stripe webhook and then auto submits evidence to try win them, it even includes the interior designs they generated in the evidence PDF to prove they used it! Here's the prompt/skill I made: ---- Build an auto-dispute-response system for Stripe that: 1. Shared evidence collection (app/dispute_evidence.php) Create a shared file with functions used by
vibe_codingautomationstripeai_assisted_developmentskills_sharing
Social Twitter Apr 6

There were likely no major work impacts of GenAI in any large firm throughout 2025. We did not have ...

By @emollick

72 score
AI Analysis

Ethan Mollick argues there were likely no major work impacts of GenAI in large firms throughout 2025, as agentic tools weren't available and adoption takes time. Cautions that 2025 studies don't predict 2027 impact.

There were likely no major work impacts of GenAI in any large firm throughout 2025. We did not have agentic tools, adoption takes time, and everyone was experimenting with process. That is starting to change. Studies that show no impact in 2025 don't tell us much about 2027.
GenAI-enterprise-adoptionAI-economic-impactagentic-toolsAI-hype-reality
Social Twitter Apr 6

Every author who sued OpenAI just got the smoking gun they needed. AI companies told courts their ...

By @AlphaSignalAI

72 score
AI Analysis

AlphaSignal reports on 'Alignment Whack-a-Mole' paper showing GPT-4o, Gemini, and DeepSeek reproduce up to 90% of copyrighted books verbatim after fine-tuning, with 90% overlap across providers. Fine-tuning on one author unlocked text from 30+ unrelated authors.

Every author who sued OpenAI just got the smoking gun they needed. AI companies told courts their models don't store copyrighted books. A new paper just proved they do. Researchers fine-tuned GPT-4o, Gemini, and DeepSeek on a simple task. Expand plot summaries into full text. No jailbreaks. No tricks. The models started reproducing entire copyrighted novels word-for-word. Up to 90% of full books. Single passages over 460 words long. The wildest part. They fine-tuned only on Murakami n
copyrightai_legalmemorizationsafety_filtersresearch_paper
Social Twitter Apr 6

We’re very excited to deepen our work with the @SAIRfoundation co-founded by Terence Tao. We’ve been...

By @Thom_Wolf

72 score
AI Analysis

Thomas Wolf (Hugging Face co-founder) announces deepening collaboration with the SAIR Foundation co-founded by Terence Tao, working on AI x science in chemistry, physics, biology, and math. Focus on sharing datasets, building challenges, and communities for open science.

We’re very excited to deepen our work with the @SAIRfoundation co-founded by Terence Tao. We’ve been very active pushing the communities in AI x science in chemistry, physics, biology (more on that very soon) and this aligns perfectly with what the SAIR foundation has been doing in math (and soon extending as well). Sharing datasets, building challenges and communities Exciting future for open-science
ai_for_scienceopen_sciencehugging_facecollaboration