R[AI]DAR

Category intelligence

Social Media Briefing

All 400 current items, analyzed and ranked.

Executive synthesis

Social Media Summary

AI progress and the open-versus-closed debate dominated discussions. Ethan Mollick argued meaningfully better AI releases are accelerating from OpenAI and Anthropic, backed by a model timeline. He separately cautioned that open-weights models are more fragile out-of-distribution than benchmarks suggest.

Note: Near-duplicate posts (Mollick acceleration, Gebru ethics-rhetoric, and vLLM thread fragments) were consolidated to a single representative each.

Key Themes

AI economics and valuation skepticism · 14AI progress and benchmarking · 5Inference Infrastructure & Serving · 6Model Behavior Comparison · 1Model Progress and Behavior · 6Open vs Closed Models · 2Agent and Developer Tooling · 5AI Industry Critique and Economics · 4Tesla, xAI and Musk criticism · 16AI safety and open science · 4

Primary evidence

All Ranked Signals

Showing 10 of 400 signals 2 remaining
Social Twitter May 30

It does seem like meaningfully better AI releases are accelerating, especially from OpenAI & Ant...

By @emollick

72 score
AI Analysis

Emollick argues that meaningfully better AI releases are accelerating, especially from OpenAI and Anthropic, sharing a timeline of models that gained 3+ points on the Artificial Analysis index.

Source details
It does seem like meaningfully better AI releases are accelerating, especially from OpenAI & Anthropic. To illustrate, I caused this timeline to be created. It only lists new models that scored 3 points or higher over previous models in the Artificial Analysis index. t.co/rxc5f3FAsO
AI progressbenchmarkingOpenAIAnthropic
Social Twitter May 30

I think Epoch does a great job benchmarking, but I continue to believe that open weights models are ...

By @emollick

72 score
AI Analysis

Emollick praises Epoch's benchmarking but argues open weights models are more fragile out-of-distribution than benchmarks suggest, doubting they were only a few months behind frontier models.

Source details
I think Epoch does a great job benchmarking, but I continue to believe that open weights models are much more fragile, especially out-of-distribution, than their benchmarks indicate. Vibe-wise, I don’t think they were only 3 months behind last year or only 4 months behind today.
open weightsbenchmarkingAI progress
Social Bluesky May 30

It does seem like meaningfully better AI releases are accelerating, especially from OpenAI & Anthrop...

By @emollick.bsky.social

72 score
AI Analysis

Ethan Mollick observes that meaningfully better AI releases are accelerating, especially from OpenAI and Anthropic, illustrated by a timeline of models scoring 3+ points above predecessors on the Artificial Analysis index.

Source details
It does seem like meaningfully better AI releases are accelerating, especially from OpenAI & Anthropic. To illustrate, I caused this timeline to be created. It only lists new models that scored 3 points or higher over previous models in the Artificial Analysis index.
model progressbenchmarksOpenAIAnthropic
Social Twitter May 30

vLLM v0.22.0 is out! 459 commits from 230 contributors (63 new). 🎉 Highlights: DeepSeek V4 hardenin...

By @vllm_project

70 score
AI Analysis

vLLM announces v0.22.0 release with 459 commits from 230 contributors, highlighting DeepSeek V4 hardening, Rust frontend, batch-invariant FP8 latency gains, and KV cache offloading.

Source details
vLLM v0.22.0 is out! 459 commits from 230 contributors (63 new). 🎉 Highlights: DeepSeek V4 hardening (NVFP4 fused MoE, full+piecewise CUDA graph, ROCm support), experimental Rust frontend in-Tree, batch-invariant Cutlass FP8 (28.9% lower e2e latency), Model Runner V2 advances, multi-tier KV cache offloading. Thread 👇
vLLMopen source releaseinference infrastructure
Social Twitter May 30

There's a better way to serve your inference stack, you just haven't found it yet. DynoSim is a wor...

By @NVIDIAAI

68 score
AI Analysis

NVIDIA introduces DynoSim, a Rust-based workload-driven simulation of the Dynamo inference serving stack that screens thousands of deployment configurations in simulation 1,500x faster than real time before validating on hardware.

Source details
There's a better way to serve your inference stack, you just haven't found it yet. DynoSim is a workload-driven simulation of the Dynamo serving stack that turns exhaustive deployment search into a simulate-then-verify loop. Instead of testing every deployment choice, teams can model the whole stack on one virtual timeline, screen thousands of configurations in high fidelity simulation, then validate only the best candidates on real hardware. And because it's a full Rust implementation, it
inference optimizationserving infrastructureNVIDIA
Social Twitter May 30

I'd go further and say most of open science defines a lot of how AI is discussed (e.g. our Tulu 3 pr...

By @natolambert

68 score
AI Analysis

Nathan Lambert argues open science shapes how AI is discussed, citing his Tulu 3 project coining RLVR, and emphasizes the value of publicly establishing methods to reduce future noise.

Source details
I'd go further and say most of open science defines a lot of how AI is discussed (e.g. our Tulu 3 project coining RLVR). There's a lot of value in this sort of level setting and cutting out future noise in science by establishing methods publicly.
open scienceRLVRAI research culture
Social Twitter May 30

AI for accelerating research, by expanding what mathematicians and scientists dare attempt:

By @gdb

66 score
AI Analysis

Continuing the Social thread on AI accelerating research, Brockman highlights AI for accelerating research by expanding what mathematicians and scientists dare attempt.

Source details
AI for accelerating research, by expanding what mathematicians and scientists dare attempt:
AI for scienceAI progressOpenAI
Social Twitter May 30

Given that Claude seems so lazy in chat (especially with technical search topics), it seems pretty t...

By @natolambert

65 score
AI Analysis

Nathan Lambert observes Claude seems lazy in chat search while GPT-5.5 and OpenAI models are more thorough, arguing the Codex harness is a lighter intervention yet Claude Code does excellent agentic tasks like extracting paper figures into slides.

Source details
Given that Claude seems so lazy in chat (especially with technical search topics), it seems pretty telling about how a harness can make a model far more independent and thorough. GPT 5.5, and many of OpenAI's recent models, seem incredibly thorough -- like they won't give up -- and the codex harness is a much lighter change on the model. Of course I have a lot of uncertainty here, but it's surprising to me how weak Claude's search is when I try the Claude app again. I only use ChatGPT for rese
model comparisonagent harnessesClaude vs GPT
Social Twitter May 30

AI safety can't happen behind closed doors! Super cool to see that the @AISecurityInst is releasing ...

By @ClementDelangue

62 score
AI Analysis

Hugging Face CEO Delangue praises the AI Security Institute for openly releasing its evals, datasets, and models on Hugging Face so researchers can scrutinize and reproduce them.

Source details
AI safety can't happen behind closed doors! Super cool to see that the @AISecurityInst is releasing its evals, datasets, and models in the open on @huggingface, so researchers everywhere can scrutinize, reproduce, and build on them: t.co/pLaKgl0oH1 t.co/yV0XeSyage
AI safetyopen sourceevaluations
Social Twitter May 30

The debate on if open or closed models win comes down to if there is disproportionate value to margi...

By @natolambert

62 score
AI Analysis

Nathan Lambert frames the open-versus-closed model debate around whether marginally better intelligence carries disproportionate value, predicting closed models stay slightly smarter while open models stay cheaper.

Source details
The debate on if open or closed models win comes down to if there is disproportionate value to marginally better intelligence. The believers of this sit across from the open models will be good enough camp. Closed models will stay slightly smarter. Open models will be cheaper.
open vs closed modelsAI strategymodel economics