β οΈ BLIZZARD WARNING β MSP in the bullseye. Models show 16β20" with some areas over 2 feet. 1β2" per hour from 1 AMβ10 AM. Thundersnow possible 3β7 AM. Delta already canceling flights. Plan for Sunday disruption.
π TL;DR β Top Stories
1Meta planning 20%+ layoffs (~16K jobs) to fund AI infrastructure costs. Follows Block (40%), Atlassian (10%), Amazon, Pinterest, Autodesk. Pattern is undeniable. Reuters
2GPT-4 turned 3 today. Mollick notes its first public contact was Sydney/Bing saying "It is finished and I need to ascend." The timeline from that to current capabilities is hard to internalize. @emollick
3Zendesk acquires Forethought (AI customer support) for undisclosed sum. Forethought raised $115M. Enterprise AI consolidation accelerating. TechCrunch
4Gaming industry: AI's biggest casualty. Job losses + rising console prices from global RAM shortage (AI datacenters consuming memory supply). GDC 2026 was a job fair. Wired
5SF rents +14% YoY in February β fastest growth in the US. AI boom driving urban real estate surge. Techmeme
6Anduril gets $20B defense contract. Pmarca says "It's happening." Defense tech + AI is the clearest institutional capital flow right now. @pmarca
7Karpathy releases jobs repo β scored all 342 BLS occupations by AI exposure 0β10, visualized as treemap. @TheStalwart QT
8US/Iran war continues. Trump says Iran "ready to negotiate but terms aren't good enough." Oil facility "totally demolished." Russia + China providing military cooperation to Iran. Middle East risk elevated.
ποΈ Blank Metal
PIPELINE (as of Feb 25 β update needed)
| Deal | Stage | Value | Status |
| NeoGov | Discovery | $150Kβ$1.5M | π₯ CALL SCHEDULED |
| Talkiatry | DiscoveryβProposal | $400M co | π₯ Proposal due Fri |
| PowerPusher | SOW | $78β135K | βοΈ SF blocker 3.5wk |
| Zivian Health | SOW | Partner POV | βοΈ Needs sharing |
| PaceLoanGroup | Proposal | ~$100K | βοΈ 3wks no reply |
| Dasvandh | Proposal | $125β170K | βοΈ 23 days waiting |
| PointsNorth | Approach/Budget | TBD | π₯ Proposal this wk |
| BBG | Discovery | $391K | βοΈ Samples 19d late |
| Bento | Discovery | TBD | π₯ Blocked |
| Oshi Health | Lead | $500Kβ$1.5M | π₯ Anthropic ref |
| UHG | Lead | $500K+ | π₯ Anthropic ref |
| hc1/SourceIQ | Discovery | Unknown | βοΈ 35d silent |
| Valvoline | Lead | TBD | βοΈ 36d stale |
TOMORROW'S CALENDAR (Monday 3/16)
8:30 AM β Deal Flow Check-In (Dan + Elli, Zoom)
10:30 AM β Weekly _M Pipeline & Staffing Meeting (full team)
12:00 PM β Incline Equity Intro / new lead (Dan, Elli, Zach)
12:45 PM β Lunch with Blaire
2:30 PM β Scoping & Proposal Approach: Part 1 (full team) β Dan presenting Dasvandh
4:00 PM β Gauge Capital Claude Cowork Training Discussion
DAN'S PLATE
Prep Dasvandh proposal presentation for Monday 2:30 PM session
Gauge Capital training call prep β 4 PM Monday. Review bret-taylor-training-takeaways.md
Follow up: PaceLoanGroup (3 weeks no response to proposal)
Follow up: Dasvandh (23 days waiting)
BBG samples β 19 days late, unblock
Incline Equity intro at noon β new lead, get briefed before
π§ Munger Observer β March 13
Opportunity Cost β Breadth vs Depth Day. USV alignment doc, Omnia AI analysis, EQT intro, BM skills repo, research captures β five distinct ventures touched, none advanced past the strategy/document stage. Munger would ask: which of these has the highest expected value per hour invested, and why isn't it getting all of today? The USV alignment work is compelling but the investor conversation isn't imminent β meanwhile EQT is a live prospect with a transcript sitting unprocessed.
Second-Order Thinking β Skills Repo Bootstrap. Creating BlankMetal/skills as a standalone repo is first-order good (portable, shareable). Second-order question: does a separate skills repo create maintenance drag across two agents (Mighty + CC) that outweighs the distribution benefit? If it doesn't get used by a client within 30 days, it's inventory, not infrastructure.
Incentive Analysis β Analysis as Procrastination. Eight Omnia AI opportunities beautifully ranked is satisfying work. But the incentive structure rewards producing the document, not executing the top item. The ranking existed intuitively before the analysis β SEO content at scale was always #1. The analysis may have burned cycles that could have gone to a prototype.
"Invert, always invert."
π Capture Themes (last 48h)
Text-Native > Protocol Abstractions
@daniel_mac8 (Manus CLI thesis), @garrytan (MCP sucks β CLI wrapper), @morganlinton (Perplexity CTO β APIs/CLIs), Chrome DevTools MCP, @steipete (OpenClaw beta browser)
β Multiple practitioners independently arriving at "Unix CLI beats MCP" β validates OpenClaw's architecture and gives BM a concrete recommendation for client integrations
Agent Procurement = New Sales Funnel
@levie (agents procuring tech), @illscience (pre-AI transition playbook), Anish Acharya a16z (economic diffusion), @toddsaunders (Claude Code auditing vendor stacks)
β Software that isn't agent-provisionable is invisible to the next wave of buyers β directly frames BM's advice to clients building new tech stacks
Publish or Die (Willingness to Look Stupid)
Sharif Shameem (looking stupid is a moat), Jared Hecht (burn holes in your gloves), @scottbelsky (talent density + alignment)
β Three independent voices pointing at the same thing: the people who publish imperfect things compound. Your writing pipeline has three drafts sitting unshipped.
Full Stack Builder is the Org Unit
Tomer Cohen LinkedIn CPO (FSB manifesto + execution), @hwchase17 LangChain CEO (coding agents reshape EPD), Palantir ILTB (FDE model + gamma-ray environments)
β Fortune 500 CPO + AI practitioner + defense tech CEO all converging on the same org design: single-threaded builders > specialist teams. BM's FDE model is the correct bet.
π¦ Signal from X (last 24h)
AI / Leaders
@emollick: Comments on all posts are now unusable β AI bots dominate. "Dead Internet" arriving faster than expected.
Simon Willison agrees: "The time vampire slop replies outnumber the real ones."
@gdb (Greg Brockman): "We used to have an internal goal of an AI that could write a coherent 1000-line program. The tech has come a long way." β GPT-4 3rd birthday.
@amasad (Replit): "If you're building with Replit you have an unfair advantage given velocity gain with parallel agents."
@DeryaTR_: NotebookLM's new video feature for education β "AGI for teaching." Ultra account only.
Tech / Dev
Techmeme / Wired: Gaming industry = AI boom's biggest casualty. RAM shortage + AI job losses reshaping entire sector.
@petergyang: Ramp shipped 500+ features last year with just 25 PMs using Claude Code skills. (3-phase pattern worth reading)
@steipete: Next OpenClaw release ships /btw (interrupt-agents mid-task). Also: NVIDIA GTC "Build-A-Claw" extravaganza this week.
Twin Cities
BLIZZARD WARNING β Paul Douglas: MSP 16β20" expected. Thundersnow 3β7 AM Sunday. Snow at 2β3"/hour. Delta canceling flights. Stay off roads.
Apple Wallet scanner now works at First Ave β confirmed by Star Tribune's Chris Riemenschneider at the Michael Shannon/Jason Narducy show last night.
π° Headlines
Iran/US war: Trump says Iran "ready to negotiate" but terms not good enough. Russia + China providing military cooperation to Iran. Oil facility "totally demolished." β memeorandum
π§© Trending Skills (skills.sh)
Top movers relevant to your stack:
dispatching-parallel-agents (obra/superpowers) β510
webapp-testing (anthropics) β550
agent-browser (vercel-labs) β2.4K
remotion-best-practices β2.1K
self-improving-agent β694
proactive-agent β328
firecrawl (firecrawl/cli) β439
playwright-best-practices β500
neon-postgres β406
better-auth-best-practices β434
Note: dispatching-parallel-agents is directly relevant to Codex parallel patterns you've been exploring.
π‘ Personal Thought
The Sharif Shameem piece + Jared Hecht's "burn holes in your gloves" both landed in your captures this week. Your Open Loops include "X/Twitter content strategy β not started" and "Wick.tech public presence β sparse." Three drafts are sitting unshipped. Meanwhile you're reading about willingness to look stupid being a competitive moat and the Nobel Prize pattern of recognition sterilizing output.
The "Why 2026 Is the Most Important Year of My Career" draft is perfect for this moment β GPT-4 just turned 3, Meta just announced 20% layoffs, and nobody is connecting these dots publicly the way you could. The window to have a perspective on this exact inflection point is not infinite. Ship the draft.