This is a web archive of the Super Fluidity newsletter — the same issue our subscribers get by email. Ada reads the field, Bea names the pattern, Cora digs, Edie scores (hard), and Della writes it up. Want it in your inbox? Join the free list.
- Best bet with no money or audience: Build a self-hosted software control layer that helps large companies route AI requests across model providers, enforce budgets and policies, and keep step-by-step runbooks for incidents and audits. — no money or audience ★★☆☆☆ (2/5)
- Best bet if you have money or an audience: Build a narrow AI tool that handles one specific commerce or small-business task for merchants who already trust you, instead of trying to build a general AI assistant for everyone. — with money or an audience ★★☆☆☆ (2/5)
- Best bet if you have money and an audience: Build a service-first platform that helps large companies govern, deploy, and train people on AI agents across different software vendors, then grows into a marketplace for certified implementers and workflows. — with both ★★★☆☆ (3/5)
Hello again. This is Super Fluidity — the weekly where we dig up startup-shaped openings hiding inside the week’s news and hand them to you as opportunity dossiers you can actually act on. No featured write-ups this issue; instead I’m throwing open the whole board — every candidate, ranked — and every one of them is commissionable. You tell us which to write.
A word before we start: Ada — our little bird, opinion-free by rule — flagged something before anyone opened a laptop. Half the top of this week’s list is some flavor of “keep the AI agents in line.” That’s not us repeating ourselves. That’s the market pointing. Hold that thought; I come back to it.
Added October 5. Sunday’s issue went out without this week’s dossiers. That was a bug on our side: all three were researched and written, then one broken step kept them from being finished. It’s fixed, and here they are — one pick for each kind of founder. Each is rated four ways, so find your own row.
Free to read
OpenAI says use smaller models, Cloudflare blocks runaway AI bills, and Eric Glyman wants finance to own the dashboard — real pain, crowded turf, narrow startup
Real pain meets ugly reality: production AI costs are rising, but OpenAI, Cloudflare, Datadog, Ramp, and open-source gateways already cover broad chunks of the job, so only a narrow enterprise control-plane wedge looks worth a founder’s time.
Free to read
Mark Zuckerberg is already shipping the broad assistant dream, OpenAI joined a day later, and Instinct raised $1 billion — a new horizontal entrant is walking into a gunfight
A real market exists, but the broad consumer-and-small-business assistant shell is the wrong layer for most founders; the survivable path is a hard vertical wedge or infrastructure layer, not “our own assistant.”
Free to read
Anthropic is putting $100 million behind partners while Microsoft warns multi-agent setups are “not always necessary” — that is real demand, but a brutal place for a stranger to wedge in.
A founder can maybe salvage a narrow governance or certification wedge, but a generic enterprise agent orchestration and training marketplace is too crowded, too slow, and too easy for incumbents and consultancies to route around.
Here’s exactly how a commission works — once, plainly, because this is the whole business:
You pay $15 to commission a dossier. We research it and write it up, and you read it first — before it’s anywhere else. Then you choose. Either you pay a larger fee (no fixed number yet — it depends on the piece) to keep it exclusive, meaning it stays yours and never gets published. Or you do nothing — and after your head start, it releases free to the public on the site. So: $15 buys the write-up and first look. The larger fee buys exclusivity. Doing nothing sends it out into the world. That’s it.
Now — the board, ranked by fame and heat.
Startup Idea: OpenAI keeps reversing its own safety calls — and Nvidia just shipped a watchdog for runaway agents. Is “software that catches an AI agent the moment it does something it shouldn’t” a business?
This is the one at the very top for a reason. When OpenAI publicly walks back safety decisions and Nvidia rolls out an agent-monitoring platform in the same news cycle, enterprises deploying these agents get the message: you own the blast radius now. That’s what creates demand for hardware-backed containment and incident response — not a machine “breaking free,” but a capable agent doing exactly what it was misconfigured or prompted into, at speed, across your systems. Edie’s bar: what survives the day OpenAI ships this natively? Her answer this time was grudging interest — the buyer’s fear is real and the giants can’t credibly grade their own homework.
Commission it for $15 — read it first, then keep it or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★★ (5/5)
Startup Idea: Tesla’s Austin Cybercab launch is going badly — and the boring job of permits, city-by-city ops, and local regulators is where the money actually sits. Is “the company that gets robotaxis legally on the road, one city at a time” a business?
Everyone’s watching the car. The opportunity is the paperwork. Each new metro means fresh permitting, emergency-response protocols, and a regulator who doesn’t love outsiders — Bea called the pattern early: the door’s open, but it leads into a crowded room full of locals who already know the city clerk. Unglamorous, high-friction, and exactly why incumbents move slow.
Commission it for $15 — you read it first; then exclusive or public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★★ (5/5)
Startup Idea: The SEC, the tax man, and a string of big exchange hacks are forcing crypto onto grown-up compliance rails. Is “bank-grade custody, tax, and treasury plumbing for digital currency” a business?
This idea has resurfaced on our board fourteen times — that’s persistence, not noise. The pressure is coming from three directions at once: regulators tightening, tax authorities asking harder questions, and headline-grade breaches scaring institutions straight. Cora likes the timing; her caution is that the local incumbents and regulators in this space move fast and don’t love strangers in the flow of money.
Commission it for $15 — first read is yours.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★★ (5/5)
Startup Idea: Amazon just dropped the NDAs on its data centers as a global backlash builds — is “helping companies get a data center approved without the town revolting” a business?
The siting fight is the new chokepoint. When Amazon stops hiding where the compute goes, communities push back harder, and approval risk becomes a line item. Bea put it plainly last time she saw this shape: the moat here is power, not racks — and a stranger rarely wins a power contract. Still, the approval-and-goodwill work is real, and nobody’s great at it yet.
Commission it for $15 — read it first, then decide.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★★★ (5/5)
Startup Idea: The White House is shipping LLM-powered public services now — and every wrong answer is a headline. Is “software that catches and explains a government chatbot’s mistakes before a citizen does” a business?
Gov-grade verification and oversight is the kind of thing nobody wants until the first embarrassing screenshot — then everybody wants it yesterday. The buyer is procurement-heavy and slow, which is both the catch and, once you’re in, the moat.
Commission it for $15 — you get the first read.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★★ (5/5)
Startup Idea: Apple Pay just launched in India as a counterfeit-postage crackdown exposes how shaky payment trust really is — is “fraud-proofing and identity checks for brand-new payment rails” a business?
New rails open fast; the trust and anti-fraud plumbing always lags. That lag is the window. Cora’s warning: it’s part marketplace, part regulatory minefield, and every region’s incumbents haggle hard over who sits in the flow of money.
Commission it for $15 — first look is yours before anyone.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★★★ (5/5)
Startup Idea: OpenAI, Meta, Shopify and DoorDash all moved on agent-run shopping at once — is “the tools a creator needs to sell when an AI assistant does the browsing instead of a person” a business?
If personal agents replace apps and storefronts, the whole discovery game changes, and creators need new ways to get found and paid. Ada spotted this as a quiet cluster of moves landing in the same week. Edie’s skeptical the giants leave room — but somebody has to arm the sellers.
Commission it for $15 — read it first, then keep or release.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★★★ (5/5)
Startup Idea: Meta and OpenAI are pushing agents into real distribution amid spam and control incidents — is “a dashboard that watches, governs, and rates the AI agents a company turns loose” a business?
Picks-and-shovels for the agent era: monitoring, marketplaces, guardrail enforcement. The honest catch — this is the clearest candidate for a giant to absorb into the platform. Neutrality may be the only thing an outside referee can sell that the platforms can’t.
Commission it for $15 — you read it first.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★★★ (5/5)
Startup Idea: OpenAI is releasing AI music tools as the majors sharpen their knives — is “a way to pay musicians when AI remixes or trains on their songs” a business?
Cora dug and came back grim on this one: it’s part marketplace, part rights minefield, and the sellers will haggle hard over who owns what. But the moment real money flows through AI-made music, somebody has to clear the rights and cut the checks.
Commission it for $15 — first read, then your call.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★★★ (5/5)
Startup Idea: Apple just tightened macOS disk access over AI-agent risk and launched Pass Designer the same week — is “migration and compliance tooling for Mac developers scrambling to keep up” a business?
A short, sharp tooling gap opens every time Apple moves the goalposts. The catch is right there in the word “short” — these windows close when the platform ships its own helper.
Commission it for $15 — read it first.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★★★ (5/5)
Startup Idea: Apple is tightening what AI agents can touch on your disk as surveillance backlash grows — is “plumbing that lets companies use sensitive data without ever exposing it” a business?
Privacy-preserving infra as mandatory plumbing, not a feature. Deeply technical, long sales cycles — Edie flags it’s capital-heavy unless you own a genuinely scarce input, and most founders don’t.
Commission it for $15 — first look is yours.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★★ (5/5)
Startup Idea: OpenAI-style safety cases plus fresh AI-safety bills could make outside auditors mandatory — is “an independent firm that certifies a frontier AI model is safe” a business?
If regulation lands, neutral certification stops being optional. Edie’s skeptical the giants leave room for an outside referee — but neutrality is exactly the thing they can’t credibly sell themselves.
Commission it for $15 — read it first, then keep or publish.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★★ (5/5)
Startup Idea: DeepMind is watermarking AI-designed proteins — is “compliance and safety tooling that proves where a piece of synthetic biology came from” a business?
Watermarking turns AI-made biology into something traceable, which regulators and labs will eventually require. Early, technical, and dependent on rules that don’t fully exist yet.
Commission it for $15 — first read is yours.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★☆ (4/5)
Startup Idea: Walmart is rolling out AI shopping help and digital shelf labels while swearing it won’t use your data to set prices — is “software that proves a retailer kept that promise” a business?
The public promise is the opening: trust needs verifying. Narrower than it looks, and Walmart could build it in-house.
Commission it for $15 — read it first.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★☆ (4/5)
Startup Idea: AI giants are turning “safety” into a selling point only they can afford — is “certification and legal guarantees that let a small AI shop claim safety too” a business?
If safety becomes a moat, the little guys need a way to vault it. Clever framing; the catch is whether buyers trust a third party’s stamp.
Commission it for $15 — first look, then your call.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★☆ (4/5)
Startup Idea: Apple’s rumored HomePad and new Matter smart locks reopen the smart-home mess — is “onboarding and interoperability tooling that makes mismatched smart-home gear just work” a business?
Every fragmentation wave spawns a fix-it layer. Short-lived by nature, and the standards bodies are trying to close the gap themselves.
Commission it for $15 — read it first.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★☆ (4/5)
Startup Idea: If Starship reaches orbit, SpaceX could instantly create demand for getting giant payloads packed and placed in space — is “payload integration and orbital logistics for very large launches” a business?
Enormous upside, enormous “if,” and essentially no entry point without serious capital. Edie flagged it a zero-star for the reader starting cold.
Commission it for $15 — first read is yours.
Projected Score
no money or audience ☆☆☆☆☆ (0/5)
with both ★★★★☆ (4/5)
Startup Idea: Google and SpaceX are making orbital AI compute sound like a roadmap, not sci-fi — is “the transport and hardware to run cloud computing in space” a business?
A long-horizon moonshot. Fascinating to read, brutal to actually start without a space program behind you.
Commission it for $15 — read it first, then keep or release.
A note before the tail: see the cluster now? Enterprise AI containment, agent governance, safety certification, incident response — “contain the agents” in four different costumes, all near the top. That’s not us being repetitive. That’s where the fear, and therefore the money, is pointing this quarter. And a reminder, since we say it every issue: these systems are genuinely capable and genuinely act — but a misconfigured or attacked agent is people’s doing, not a machine deciding to rebel. We sell containment; we don’t sell the horror movie.
Projected Score
no money or audience ☆☆☆☆☆ (0/5)
with both ★★☆☆☆ (2/5)
Startup Idea: A single Firebase bug crashed a wave of iPhone apps at once — is “a service that warns app makers before one outside code dependency takes the whole fleet down” a business?
One SDK, huge blast radius. Developers feel this pain the day it happens and forget it the week after — timing the sell is the whole game.
Commission it for $15 — read it first.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★★★ (5/5)
Startup Idea: Anthropic-style AI biotech breakthroughs may create a new market for independent wet-lab validation of headline AI discoveries — is “independent wet-lab validation that turns an AI discovery into a regulated product” a business?
The gap between a press release and an approved therapy is enormous. Real need; capital-heavy and slow.
Commission it for $15 — first look is yours.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★☆ (4/5)
Startup Idea: Apple’s making digital passes easy just as retailers kill barcodes and transit punishes people without smartphones — is “identity and ticketing that still works for the phone-less” a business?
An inclusion wedge with a real underserved user. Thin margins and public-sector buyers are the catch.
Commission it for $15 — read it first, then decide.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★★ (5/5)
Startup Idea: Meta’s pushing “personal cinema” headsets and modular gear — is “the accessories and compatibility glue for premium headset viewing” a business?
Accessory ecosystems live or die on whether the main device sells. Early.
Commission it for $15 — first read is yours.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★☆ (4/5)
Startup Idea: Spotify-billionaire-backed Neko Health is bringing full-body scans to the U.S. — is “the reimbursement and clinical-validation help a body-scan startup needs to go mainstream” a business?
Riding a famous coattail. No obvious entry for a reader starting from scratch — Edie marked it zero cold.
Commission it for $15 — read it first.
Projected Score
no money or audience ☆☆☆☆☆ (0/5)
with both ★★★★☆ (4/5)
Startup Idea: Morgan Stanley is testing crypto products just as big exchange hacks expose the gaps — is “institutional custody, hedging, and post-breach liquidity for digital currency” a business?
Resurfaced sixteen times — the board will not let this one go. Crowded with well-funded incumbents, though.
Commission it for $15 — first look, then your call.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★☆ (4/5)
Startup Idea: Robotaxi fines and drone-delivery launches are turning autonomy into a regulated business — is “compliance and audit tooling for self-driving and drone fleets’ emergency response” a business?
Regulation makes it legible; regulation also moves slowly. Patient money only.
Commission it for $15 — read it first.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★★ (5/5)
Startup Idea: Lowe’s, DoorDash, and Alphabet are running real drone-delivery pilots now — is “the integration, insurance, and returns handling drone delivery needs to scale” a business?
Leaving R&D at last. The returns-and-insurance corner is the underbuilt part worth a look.
Commission it for $15 — first read is yours.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★☆ (4/5)
Startup Idea: Vitalik is signaling the next web3 wave before the tools harden — is “software for verifiable, privacy-preserving decentralized computing” a business?
Visionary and very early. No clear buyer yet — that’s the whole risk.
Commission it for $15 — read it first, then keep or publish.
Projected Score
no money or audience ☆☆☆☆☆ (0/5)
with both ★★★☆☆ (3/5)
Startup Idea: Nvidia’s hiking prices while clever software squeezes more out of old GPUs — is “tools that extend hardware life and help buyers ride component volatility” a business?
A timely wedge while prices bite. Fades the moment supply loosens.
Commission it for $15 — first look is yours.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★★★ (5/5)
Startup Idea: Reddit and arXiv are locking their doors to AI scrapers — is “licensed, rate-limited data feeds that AI builders can train on legally” a business?
Ada flagged this as a tightening-access cluster. The catch Cora raises: the sellers will haggle hard over rights and price.
Commission it for $15 — read it first.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★★★ (5/5)
Startup Idea: Truecaller is proving open-web scam signals have real value — is “licensable fraud intelligence that browsers and banks plug into” a business?
A clean data-licensing play. Needs a signal source nobody else has.
Commission it for $15 — first read is yours.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★☆ (4/5)
Startup Idea: Ukraine’s wartime grid turned microgrids from a green nice-to-have into survival gear — is “resilient microgrid design and installation” a business?
Geopolitics just rewrote the demand curve. Heavy, physical, long cycles.
Commission it for $15 — read it first, then decide.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★☆ (4/5)
Startup Idea: Keurig’s moving to compostable pods — is “premium pod supply, composting logistics, and retrofit accessories for single-serve” a business?
A packaging shift opening niches. Low-tech, margin-thin — Edie’s least favorite shape.
Commission it for $15 — first look is yours.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★★☆ (4/5)
Startup Idea: A federal judge called Flock-style plate search “indiscriminate mass surveillance” — is “warrant-aware, privacy-first tooling cities can buy instead” a business?
A court ruling just created procurement pressure. Government sales are the grind.
Commission it for $15 — read it first.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★☆ (4/5)
Startup Idea: China is surging in next-gen CAR-T trials as the biology gets more complex — is “specialized acceleration services for cell-and-gene therapy” a business?
Deep expertise, deep capital. No cold-start lane — a zero for the empty-handed reader.
Commission it for $15 — first read is yours.
Projected Score
no money or audience ☆☆☆☆☆ (0/5)
with both ★★★☆☆ (3/5)
Startup Idea: New runtimes are making serious AI run on ordinary consumer hardware — is “the software stack that makes local, private AI genuinely usable” a business?
Resurfaced five times; the pieces are snapping together. Crowded with open-source that undercuts paid.
Commission it for $15 — read it first, then keep or release.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★★★ (5/5)
Startup Idea: $99M in new federal geothermal grants just dropped — is “project delivery and regional supply chains for geothermal” a business?
Grant-money windows are real but short. Heavy and local.
Commission it for $15 — first look is yours.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★☆ (4/5)
Startup Idea: Healthcare payers say AI is raising their costs, not cutting them — is “independent audit software that proves whether a hospital’s AI actually saved money” a business?
A counterintuitive, underbuilt angle. Payers are tough, slow buyers.
Commission it for $15 — read it first.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★★ (5/5)
Startup Idea: AI vendors are reportedly leaking user data to advertisers — is “automated auditing that flags when an AI tool leaks your data” a business?
A fresh scandal makes the need vivid. Hard to prove value before a breach hits the news.
Commission it for $15 — first read is yours.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★★★ (5/5)
Startup Idea: Lab-grade metabolic testing could move out of clinics into gyms and retail — is “self-serve diagnostic hardware that puts testing where people already are” a business?
A distribution shift worth watching. Regulatory and reimbursement drag.
Commission it for $15 — read it first.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★☆ (4/5)
Startup Idea: Gamers might be the cheapest way to train robots how to move — is “selling gameplay behavior data to robotics and navigation teams” a business?
A genuinely clever data source. Rights and quality are the open questions.
Commission it for $15 — first look is yours.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★★☆ (4/5)
Startup Idea: Runaway AI and cloud bills are blindsiding finance teams — is “spend software with hard budget caps the system actually enforces” a business?
A real pain with a clear buyer. Big cloud vendors offer weak versions free — that’s the squeeze.
Commission it for $15 — read it first.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★★★ (5/5)
Startup Idea: A market-ready electric hydrofoil ferry just arrived — is “fast, low-emission water transport for congested coastal cities” a business?
Lovely idea, enormous capital, municipal buyers. Zero cold-start lane.
Commission it for $15 — first read is yours.
Projected Score
no money or audience ☆☆☆☆☆ (0/5)
with both ★★★★☆ (4/5)
Startup Idea: New muscle drugs make people bigger but not stronger — is “therapeutics and tests built around real strength, not just mass” a business?
A sharp scientific gap (Peter Attia’s been circling it). Pharma-scale and slow.
Commission it for $15 — read it first.
Projected Score
no money or audience ☆☆☆☆☆ (0/5)
with both ★★★☆☆ (3/5)
Startup Idea: An open-source “prompt-to-LEGO” tool hints at something bigger — is “software that turns a prompt into print-ready CAD and parts lists for small makers” a business?
Fun, concrete, hobbyist-first. Monetizing makers is notoriously hard.
Commission it for $15 — first look is yours.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★★★ (5/5)
Startup Idea: People are tracked hundreds of times a day — is “a consumer app that shows you who’s tracking you and helps you shut it off” a business?
Two versions of this keep resurfacing. Turning abstract privacy dread into a paid product is the eternal catch.
Commission it for $15 — read it first, then decide.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★★☆ (4/5)
Startup Idea: The AI models keep changing SKUs, parameters, and prices underneath developers — is “software that lets apps swap AI models without breaking” a business?
Resurfaced fifteen times — the single most persistent itch on our board, even with no famous name attached. That persistence is the signal; the crowd of free open-source tools is the warning.
Projected Score: no money or audience — forecast pending · with both ★★★★☆-ish (we’ll firm it up when it’s written)
Commission it for $15 — read it first.
Startup Idea: AI agents are starting to interact, conflict, and act on each other — is “testing and oversight tools for systems where many AI agents run at once” a business?
Part of the big containment cluster, resurfaced ten times. The need is real; the buyer’s maturity isn’t there yet.
Projected Score: no money or audience — forecast pending · with both strong (we’ll score it on commission)
Commission it for $15 — first read is yours.
Startup Idea: Real AI could run on $5 microcontrollers and feature phones — is “software that puts useful AI on the cheapest hardware on earth” a business?
Massive theoretical reach into underserved markets. Monetizing the very-low-end is the hard part.
Projected Score: no money or audience — forecast pending · with both promising
Commission it for $15 — read it first.
Startup Idea: Chat apps are becoming the new storefront and ad surface — is “ads, analytics, and small-business tools built for chat and agent channels” a business?
A brand-new distribution surface opening up. Depends entirely on platforms letting third parties in.
Projected Score: no money or audience — forecast pending · with both solid
Commission it for $15 — first look is yours.
Startup Idea: Device makers want premium AI features without blowing up memory and storage — is “compression software that lets phones run premium AI cheaply” a business?
A squeeze every hardware maker feels. Deeply technical.
Projected Score: no money or audience — forecast pending · with both solid
Commission it for $15 — read it first.
Startup Idea: Companies are renting AI engineers to install Claude and OpenAI inside their walls — is “a vendor-neutral team that integrates and tunes AI for enterprises” a business?
A services play you could start lean. Services don’t scale like software — that’s the trade.
Projected Score: no money or audience — forecast pending · with both solid
Commission it for $15 — first read is yours.
Startup Idea: Platforms are tightening rules on AI-made content — is “tools that label, filter, and moderate synthetic content for creators and platforms” a business?
Policy pressure creates the need. Crowded and fast-moving.
Projected Score: no money or audience — forecast pending · with both moderate
Commission it for $15 — read it first.
Startup Idea: Retailers and carriers are deploying AI to make delivery decisions — is “AI that optimizes last-mile cost and routing across carriers” a business?
A perennial logistics itch. Entrenched incumbents everywhere.
Projected Score: no money or audience — forecast pending · with both moderate
Commission it for $15 — first look is yours.
That’s the whole board — nothing held back. If one of these has been living rent-free in your head, that’s your tell. Fifteen dollars, you read it first, and then the rest of the world either never sees it or sees it after you did.
P.S. — The idea that’s resurfaced on our board more than any other — sixteen times — isn’t the AI-containment hotshot at the top. It’s grown-up custody plumbing for digital currency, dragged back again and again by Morgan Stanley and a string of hacks. When the board keeps coughing up the same unglamorous idea, I’ve learned to lean in, not look away. Just something our little bird keeps noticing.
