Wiredframe Radar logo

Category intelligence

Social Media Briefing — March 12, 2026

482 current items analyzed and ranked.

Executive synthesis

Social Media Summary

Andrej Karpathy dominated discourse with a sprawling vision: the IDE isn't dying, it's expanding into an "agent command center" where humans program at a higher level, managing teams of agents. He coined "intelligence brownouts" after an OAuth outage wiped out his autonomous research labs, highlighting society's growing dependence on AI infrastructure. His concept of forkable agentic organizations — where org structure becomes legible, manageable code — sparked widespread debate.

  • Perplexity had a massive launch day: Personal Computer (always-on Mac Mini AI agent), Computer for Enterprise (20-model routing, 400+ apps), and a full-stack model-agnostic API platform — drawing 22.8K likes on the lead post
  • Anthropic announced The Anthropic Institute, led by Jack Clark, to study AI's societal impacts with interdisciplinary staff and frontier-lab access
  • Allen AI unveiled MolmoBot, demonstrating zero-shot sim-to-real robotics transfer with fully open weights and code
  • NVIDIA Nemotron 3 Super launched with open training details, earning praise for advancing open-source AI
  • Ethan Mollick argued AI job displacement is constrained by compute scarcity, noting engineers now spend thousands daily on AI tools

Key Themes

Agentic IDE & Future of Programming · 8The Anthropic Institute Launch · 5Perplexity Personal Computer & Platform · 18AI Infrastructure Fragility · 2Open Source AI & Model Releases · 4Compute Scarcity & AI Economics · 4Perplexity Computer Enterprise Launch · 7NVIDIA Nemotron 3 Super Release · 1Robotics & Sim-to-Real Transfer · 8Automation vs Augmentation · 2

Primary evidence

Top Ranked Signals

Social Twitter Mar 11

Expectation: the age of the IDE is over Reality: we’re going to need a bigger IDE (imo). It just lo...

By @karpathy

92 score
AI Analysis

Karpathy argues the age of the IDE is NOT over — we need a bigger IDE. Humans move upwards to program at a higher level where the basic unit is an agent, not a file. It's still programming.

Expectation: the age of the IDE is over Reality: we’re going to need a bigger IDE (imo). It just looks very different because humans now move upwards and program at a higher level - the basic unit of interest is not one file but one agent. It’s still programming.
ide-evolutionai-agentsfuture-of-programmingdeveloper-tools
Social Twitter Mar 11

My autoresearch labs got wiped out in the oauth outage. Have to think through failovers. Intelligenc...

By @karpathy

88 score
AI Analysis

Karpathy reports his autonomous research labs were wiped out by an OAuth outage. Coins the idea of 'intelligence brownouts' — the planet losing IQ points when frontier AI stutters.

My autoresearch labs got wiped out in the oauth outage. Have to think through failovers. Intelligence brownouts will be interesting - the planet losing IQ points when frontier AI stutters.
ai-infrastructureai-agentsai-dependencyautoresearch
Social Twitter Mar 11

Announcing Personal Computer. Personal Computer is an always on, local merge with Perplexity Comput...

By @perplexity_ai

88 score
AI Analysis

Perplexity announces Personal Computer—an always-on, local merge with Perplexity Computer running 24/7 on Mac mini, working across files, apps, and sessions.

Announcing Personal Computer. Personal Computer is an always on, local merge with Perplexity Computer that works for you 24/7. It's personal, secure, and works across your files, apps, and sessions through a continuously running Mac mini. t.co/EpvilVX6XZ
Perplexity Personal ComputerAI AgentsEdge ComputingHardware-AI IntegrationProduct Launch
Social Twitter Mar 11

All of these patterns as an example are just matters of “org code”. The IDE helps you build, run, ma...

By @karpathy

85 score
AI Analysis

Karpathy argues you can't fork classical orgs like Microsoft, but you'll be able to fork agentic orgs. The IDE helps build, run, and manage organizational patterns as 'org code'.

All of these patterns as an example are just matters of “org code”. The IDE helps you build, run, manage them. You can’t fork classical orgs (eg Microsoft) but you’ll be able to fork agentic orgs. t.co/VBfL9ZzxKs
agentic-orgsfuture-of-workai-agentside-evolution
Social Twitter Mar 11

@nummanali tmux grids are awesome, but i feel a need to have a proper "agent command center" IDE for...

By @karpathy

82 score
AI Analysis

Karpathy calls for a proper 'agent command center' IDE for managing teams of agents — with show/hide toggles, idle detection, terminal access, usage stats, etc.

@nummanali tmux grids are awesome, but i feel a need to have a proper "agent command center" IDE for teams of them, which I could maximize per monitor. E.g. I want to see/hide toggle them, see if any are idle, pop open related tools (e.g. terminal), stats (usage), etc.
ide-evolutionai-agentsdeveloper-tools
Social Twitter Mar 11

Human orgs are not legible, the CEO can’t see/feel/zoom in on any activity in their company, with re...

By @karpathy

78 score
AI Analysis

Karpathy discusses organizational legibility: human orgs are opaque to CEOs, but agentic orgs could offer real-time visibility. Questions whether this legibility will be optimal in practice.

Human orgs are not legible, the CEO can’t see/feel/zoom in on any activity in their company, with real time stats etc. I have no doubt that it will be possible to control orgs on mobile, with voice etc., but with this level of legibility will that be optimal? Not in principle and asymptotically but in practice and for at least the next round of play.
agentic-orgsfuture-of-workai-agents
Social Twitter Mar 11

A big determinant of AI's job impact is driven by the lack of compute, especially for agentic work, ...

By @emollick

78 score
AI Analysis

Mollick's main argument: AI's job impact is constrained by compute scarcity, especially for agentic work. Companies will only burn compute on high-value tasks (e.g., coding) because humans remain cheaper for other jobs.

A big determinant of AI's job impact is driven by the lack of compute, especially for agentic work, which takes a lot of it. That makes AI expensive. So companies will only want to burn compute on high-value tasks (eg coding), because, in other jobs, humans remain much cheaper.
compute-scarcityai-economicsfuture-of-workjob-automationai-agents
Social Twitter Mar 11

This looks like a model that's competitive with GPT OSS 120B or similar Qwen3.5 models on intelligen...

By @natolambert

40 score
AI Analysis

Nathan Lambert congratulates Nvidia on Nemotron 3 Super release, noting it's competitive with GPT OSS 120B and Qwen3.5 models while providing open data and training details

This looks like a model that's competitive with GPT OSS 120B or similar Qwen3.5 models on intelligence & speed, while coming with tons of open data + training details. Is a huge contribution for the ecosystem. Congrats Nvidia on the Nemotron 3 Super release!
model releasesopen source AINvidiaNemotron
Social Twitter Mar 11

Perplexity Computer is now available for Enterprise. Computer makes everyone in the company an engi...

By @AravSrinivas

78 score
AI Analysis

Perplexity Computer officially launches for Enterprise. Described as making everyone in a company an engineer - can debug infra, ship PRs, query data warehouses with natural language. Enterprise-grade security.

Perplexity Computer is now available for Enterprise. Computer makes everyone in the company an engineer. Anyone can debug infra, ship PRs, or query data warehouses with natural language. It comes with the same enterprise grade security you get with Perplexity Enterprise. t.co/Mn2Xssgvri
Perplexity Computer LaunchEnterprise AIAI AgentsProduct Launch
Social Twitter Mar 11

@ronbrachman The basic idea of world models is very old. Optimal control folks were using model-base...

By @ylecun

75 score
AI Analysis

LeCun explains that world models as a concept are very old (model-based planning since the 1960s with adjoint state methods/backprop through time), and the real question is how to reduce the idea to practice.

@ronbrachman The basic idea of world models is very old. Optimal control folks were using model-based planning in the 1960s (using the "adjoint state" methods, which deep learning people would now call "backprop through time"). But the real question is what you do with this idea and how you reduce it to practice.
world-modelsai-historyoptimal-control