Wiredframe Radar logo

Category intelligence

Social Media Briefing — March 17, 2026

510 current items analyzed and ranked.

Executive synthesis

Social Media Summary

GPT-5.4 adoption metrics dominated the day. Greg Brockman revealed staggering numbers — 5T tokens/day within a week, $1B annualized net-new revenue — while Sam Altman declared all 'hardcore builders' have switched to Codex, signaling OpenAI's strongest API launch ever.

Key Themes

GPT-5.4 Launch & Adoption Metrics · 8NVIDIA GTC Partnerships & Nemotron Coalition · 5Grok Multimodal Video Analysis · 5AI Agent Infrastructure & Developer Tools · 2Agentic AI & Browser Agents · 8OpenAI Codex Dominance · 2NVIDIA GTC 2026 · 14Gemini API Scaling & Pricing Overhaul · 16Open-Source AI Economics · 3Open Models Strategy & Ecosystem · 5

Primary evidence

Top Ranked Signals

Social Twitter Mar 16

gpt-5.4 has ramped faster than any other model we've launched in the API: within a week of launch, 5...

By @gdb

97 score
AI Analysis

Greg Brockman reveals GPT-5.4 API metrics: 5T tokens/day within first week, handling more volume than entire API one year ago, $1B annualized net-new revenue run rate.

gpt-5.4 has ramped faster than any other model we've launched in the API: within a week of launch, 5T tokens per day, handling more volume than our entire API one year ago, and reaching an annualized run rate of $1B in net-new revenue. it's a good model, try it out!
GPT-5.4OpenAI revenueAPI adoptionAI market scalebreaking news
Social Twitter Mar 16

Looking forward to building frontier open source AI models together with @Nvidia as we join the Nemo...

By @arthurmensch

82 score
AI Analysis

Mistral AI CEO Arthur Mensch announces Mistral is joining NVIDIA's Nemotron Coalition to co-develop frontier open-source AI base models.

Looking forward to building frontier open source AI models together with @Nvidia as we join the Nemotron Coalition and start training the first base models. t.co/OBDuMqECQp
NVIDIA partnershipsopen-source AINemotron CoalitionGTC announcements
Social Twitter Mar 16

Wow. Grok watched this video and made a complete list of everything it saw: https://t.co/fqC1fuwhwX...

By @Scobleizer

82 score
AI Analysis

Scobleizer is amazed that Grok watched a video and made a complete list of everything it saw, including reading every poster. Massively viral post with 55M views.

Wow. Grok watched this video and made a complete list of everything it saw: t.co/fqC1fuwhwX Do you have any idea how cool this is? It read every poster.
Grokmultimodal AIvideo understandingxAIpractical AI use cases
Social Twitter Mar 16

Should there be a Stack Overflow for AI coding agents to share learnings with each other? Last week...

By @AndrewYNg

80 score
AI Analysis

Andrew Ng announces a major update to Context Hub (chub), an open CLI tool for coding agents with 6K+ GitHub stars. New feature: agents can share feedback on documentation with each other, creating a Stack Overflow-like system for AI agents.

Should there be a Stack Overflow for AI coding agents to share learnings with each other? Last week I announced Context Hub (chub), an open CLI tool that gives coding agents up-to-date API documentation. Since then, our GitHub repo has gained over 6K stars, and we've scaled from under 100 to over 1000 API documents, thanks to community contributions and a new agentic document writer. Thank you to everyone supporting Context Hub! OpenClaw and Moltbook showed that agents can use social media bui
AI coding agentsdeveloper toolsagent infrastructureopen-sourceagent collaboration
Social Twitter Mar 16

🚀Announcing a strategic partnership with NVIDIA to co-develop frontier open-source AI models, combin...

By @MistralAI

80 score
AI Analysis

Mistral AI announces strategic partnership with NVIDIA to co-develop frontier open-source AI models, combining Mistral's model architecture with NVIDIA's compute infrastructure.

🚀Announcing a strategic partnership with NVIDIA to co-develop frontier open-source AI models, combining Mistral AI’s frontier model architecture and full-stack AI offering with NVIDIA’s leading compute infrastructure and development tools. t.co/4jqyKHHELz
NVIDIA partnershipsopen-source AIMistral AIGTC announcementsfrontier models
Social Twitter Mar 16

The Codex team are hardcore builders and it really comes through in what they create. No surprise al...

By @sama

78 score
AI Analysis

Sam Altman promotes Codex, saying all 'hardcore builders' he knows have switched to it, and shares that usage is growing very fast.

The Codex team are hardcore builders and it really comes through in what they create. No surprise all the hardcore builders I know have switched to Codex. Usage of Codex is growing very fast: t.co/lRKcNJDY8n
OpenAI CodexAI coding toolsdeveloper adoptioncompetitive dynamics
Social Twitter Mar 16

Computer can now take full control of Comet to complete tasks. When you’re in Comet, Computer spins...

By @perplexity_ai

78 score
AI Analysis

Perplexity officially announces Computer feature for Comet browser: a browser agent that takes full control of the browser, accesses any logged-in site without connectors or MCPs, available to all Computer users.

Computer can now take full control of Comet to complete tasks. When you’re in Comet, Computer spins up a browser agent that can access any site or logged‑in app with your permission, without the need for connectors or MCPs. Available to all Computer users on Comet. t.co/MOlJKacJia
perplexitybrowser_agentsagentic_aiproduct_launch
Social Twitter Mar 16

They tell you training and running AI model costs billions. That's true for a few frontier labs. Bu...

By @ClementDelangue

75 score
AI Analysis

HuggingFace CEO Clement Delangue presents data on real-world AI training costs, arguing most use cases are dramatically cheaper than frontier model training, advocating for smaller targeted open-source models.

They tell you training and running AI model costs billions. That's true for a few frontier labs. But for most real-world use cases? Dramatically lower than you think thanks to open-source. Real examples from @HuggingFace's latest analysis:
  • Fine-tune a text classification model: <$2k
  • Train a leading image embedding model: < $7k
  • Train Deepseek OCR: < $100k
  • Train a leading machine translation model: <$500k
Compare that to GPT-4.5 training (~$300M est.) And the truth is that you don't ne
AI training costsopen-source AIAI economicsmodel efficiencyHuggingFace
Social Twitter Mar 16

P-EAGLE from @AmazonScience and @NVIDIAAIDev removes the sequential bottleneck in speculative decodi...

By @vllm_project

75 score
AI Analysis

vLLM project announces P-EAGLE from Amazon Science and NVIDIA: parallel speculative decoding that generates all K draft tokens in a single forward pass, achieving up to 1.69x speedup over EAGLE-3 on B200 GPUs. Pre-trained heads available for GPT-OSS and Qwen3-Coder models.

P-EAGLE from @AmazonScience and @NVIDIAAIDev removes the sequential bottleneck in speculative decoding — all K draft tokens generated in a single forward pass. 📈 Up to 1.69x speedup over vanilla EAGLE-3 on NVIDIA B200, with 5-25% gains sustained at high concurrency (c=64). How it works: EAGLE drafts tokens autoregressively (K tokens = K forward passes). P-EAGLE replaces this with parallel generation using learned mask tokens and shared hidden states — one pass, K tokens. Pre-trained P-EAGLE
inference_optimizationspeculative_decodingvllmnvidiaopen_source
Social Twitter Mar 16

@ChristosTzamos Wait this is so awesome!! Both 1) the C compiler to LLM weights and 2) the logarithm...

By @karpathy

72 score
AI Analysis

Karpathy expresses excitement about a C compiler to LLM weights approach and logarithmic complexity hard-max attention with potential generalizations.

@ChristosTzamos Wait this is so awesome!! Both 1) the C compiler to LLM weights and 2) the logarithmic complexity hard-max attention and its potential generalizations. Inspiring!
attention mechanismsmodel compilationtechnical breakthroughAI research