R[AI]DAR

Category intelligence

Social Media Briefing

All 517 current items, analyzed and ranked.

Executive synthesis

Social Media Summary

The Anthropic Fable/Mythos release dominated discussion, blending excitement over capability with controversy over safeguards.

Technical and product threads also drew strong engagement:

Key Themes

Anthropic Fable and Token Economics · 6LLM Coding and Interpretability · 2Grok Build Plugin Ecosystem · 4Anthropic Fable/Mythos Controversy · 12AI Governance and Open Access · 9AI Safety and Pause Sentiment · 7OpenAI Sustainability and Bubble Concerns · 16Responsible AI and social impact · 2Agentic Search and Research · 3Sports and Real-World AI Deployment · 1

Primary evidence

All Ranked Signals

Showing 10 of 517 signals 2 remaining
Social Twitter Jun 11

It seems like LLMs could optimize coding style by exploring ways of structuring code so weaker and w...

By @ID_AA_Carmack

80 score
AI Analysis

The author speculates that LLMs could optimize coding style so weaker models can still complete tasks in a codebase, noting transformer-specific stylistic quirks likely overlap with human readability and that optimizing for at-a-glance understanding would help even frontier models.

Source details
It seems like LLMs could optimize coding style by exploring ways of structuring code so weaker and weaker models can still successfully perform tasks in a codebase. There are surely stylistic quirks that are peculiarly impactful to transformers, but I bet there would be a lot of overlap with human capabilities. Optimizing for understanding should help even the top frontier models, allowing them to understand things “at a glance” without having to explicitly explore. There will remain “better”
LLM codingcode optimizationtransformer behavior
Social Twitter Jun 11

Up until yesterday, our entire MTS team has operated under the philosophy of tokenmaxxing as much as...

By @jerryjliu0

76 score
AI Analysis

A detailed account of how the team maxed out token usage on Claude Max plans and how the Fable model is straining quotas, with one member burning the equivalent of fifteen hundred dollars in ten hours, prompting plans for model routing in engineering workflows to control burn rate.

Source details
Up until yesterday, our entire MTS team has operated under the philosophy of tokenmaxxing as much as possible on Claude Max plans. With Fable, this may no longer be possible:
  • One of our team members hit his limit 3 times yesterday and used the equivalent of $1.5k in 10 hours
  • Half of our team has hit quota limits on eng work
This era of tokenmaxxing may need to be restrained - or at least have clear guardrails defined. We are concerned about running Fable at API-based billing. If every engi
token economicsAnthropic Fablemodel routingengineering cost management
Social Twitter Jun 11

Awesome to see this innovation in text diffusion. DiffusionGemma is lightning fast, 4x faster than o...

By @demishassabis

70 score
AI Analysis

Following yesterday's News coverage of DiffusionGemma, Demis Hassabis celebrates DiffusionGemma, a text-diffusion model he says is 4x faster than other Gemma 4 models, congratulating the team.

Source details
Awesome to see this innovation in text diffusion. DiffusionGemma is lightning fast, 4x faster than other Gemma 4 models! Congrats to @bodonoghue85 and the team who worked so hard on this - excited to see what people build with it!
DiffusionGemmatext diffusionGoogle DeepMindefficiency
Social Twitter Jun 11

We're integrating Deep Research as a native skill inside Computer. It now connects to the agent har...

By @perplexity_ai

63 score
AI Analysis

Perplexity announces Deep Research as a native skill inside its Computer agent, connecting to the agent harness with search-as-code, long-running sandboxes, connectors, and licensed data, available to Pro and Max users.

Source details
We're integrating Deep Research as a native skill inside Computer. It now connects to the agent harness that powers Computer, with access to search as code generation, long running sandboxes, connectors, tools, and licensed data. Available now to Pro and Max subscribers. t.co/uHpVISkh2P
PerplexityAI agentssearchproduct launch
Social Twitter Jun 11

This tweet from 14 hours ago is on track to get about a million views. But here’s the thing: the c...

By @GaryMarcus

62 score
AI Analysis

Marcus says a WSJ scoop that OpenAI is considering drastic price cuts signals weakness, warning that an OpenAI decline could drag down Nvidia, Oracle, and CoreWeave.

Source details
This tweet from 14 hours ago is on track to get about a million views. But here’s the thing: the conclusion is true, but the tweet itself is already outdated. WSJ’s scoop that OpenAI is considering drastic price cuts - basically lighting money on fire to save customers – is further sign of weakness, and indicator of how much trouble OpenAI is in. And when they go down, they will likely pull Nvidia and Oracle and Coreweave etc down with them. Things are unraveling fast.
OpenAI criticismAI pricingsystemic riskAI bubble
Social Twitter Jun 11

Install the @sentry plugin and ask your agent to find and fix errors, analyze stack traces, and tria...

By @xai

62 score
AI Analysis

xAI promotes a Sentry plugin for Grok agents to find and fix errors, analyze stack traces, and triage alerts as part of its plugin marketplace push.

Source details
Install the @sentry plugin and ask your agent to find and fix errors, analyze stack traces, and triage alerts
coding agentsdeveloper toolsxAI
Social Twitter Jun 11

Seven challenges to the AI industry in the last 26 hours* 👉 Banks rejected SoftBank’s attempt to bo...

By @GaryMarcus

60 score
AI Analysis

Marcus lists seven recent setbacks for the AI industry, including blocked SoftBank lending against OpenAI stock, a German liability ruling against Google's LLMs, the SpaceX IPO challenge, Anthropic's Fable controversy, OpenAI price-cut considerations, market pushback on Oracle financing, and a paused Crusoe AI campus.

Source details
Seven challenges to the AI industry in the last 26 hours* 👉 Banks rejected SoftBank’s attempt to borrow again OpenAI stock 👉 Germany courts told Google that their LLMs can be held liable for untruths 👉 Senator Warren challenged the legitimacy of the SpaceX IPO 👉 Anthropic’s Fable Fiasco (already partly walked back) 👉 OpenAI has begun considering drastic price cuts, which make an already-hard to swallow profitability story harder to swallow 👉 Markets are rebelling against Oracle’s latest f
AI industry challengesAI financeregulationAI bubble
Social Twitter Jun 11

We’re launching Claude Corps, a national fellowship program matching people early in their careers w...

By @AnthropicAI

60 score
AI Analysis

Anthropic launches Claude Corps, a national fellowship pairing 1,000 early-career people with US nonprofits and paying them to use Claude to advance host missions.

Source details
We’re launching Claude Corps, a national fellowship program matching people early in their careers with US nonprofits. We'll teach 1,000 people to use Claude, and pay them to use AI to advance their hosts’ missions. t.co/QI6JmlAdSr
AnthropicAI for goodworkforcenonprofits
Social Twitter Jun 11

Use the @vercel plugin to deploy to production, spin up sandboxes, or build apps with Shadcn.

By @xai

60 score
AI Analysis

xAI promotes a Vercel plugin for Grok that can deploy to production, spin up sandboxes, and build apps with Shadcn.

Source details
Use the @vercel plugin to deploy to production, spin up sandboxes, or build apps with Shadcn.
coding agentsdeveloper toolsxAI
Social Twitter Jun 11

Deep Research in Computer is built on our Search as Code architecture. The model writes code that a...

By @perplexity_ai

60 score
AI Analysis

Perplexity describes Deep Research in Computer built on a Search as Code architecture where the model writes code to run thousands of parallel retrieval steps, claiming benchmark wins over legacy Deep Research.

Source details
Deep Research in Computer is built on our Search as Code architecture. The model writes code that assembles search itself, running thousands of retrieval steps in parallel, tailored to each question. It outperforms legacy Deep Research on every benchmark. t.co/0BrQjt8iGc
PerplexitysearchAI agentsbenchmarks