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 software that reads incoming vulnerability reports for security teams, filters obvious AI junk and duplicates, and routes the real reports into Jira or email workflows. — no money or audience ★★☆☆☆ (2/5)
- Best bet if you have money or an audience: Build a niche advisory firm that helps regulated companies prove their AI rules, policies, and controls are ready for one specific framework or deadline, then sell ongoing evidence and workflow support. — with an audience ★★★☆☆ (3/5)
- Best bet if you have money and an audience: Build software for publishers and enterprise content teams that checks AI-made text, images, audio, and video across many vendor tools, then keeps the audit trail needed for EU compliance. — with both ★★★☆☆ (3/5)
Hello again. This is Super Fluidity — the weekly where my team and I hand you opportunity dossiers: a real business someone could build, the famous move that cracked it open, and the honest reason you might want to walk away. Every item on the board below you can commission us to research and write (more on exactly how that works when we get there).
This week the news gods served up three brand-name AI stories, and in all three my job is the same: show you the juice, then show you the hook hidden under it. Edie, our resident skeptic, didn’t hand out a single star above two this week — and twice she came back MORE generous than we guessed. That almost never happens. Let me show you why.
Free to read
Google froze part of its open-source bug bounty after an AI-spam surge — and the man who shut curl’s rewards first tells you why this is a feature, not a company
Here’s the scene. Daniel Stenberg, who built curl — the tiny program running inside billions of devices — pulled the plug on curl’s bug bounty on January 26, 2026, after what he called an “explosion in AI slop reports.” His confirmed-report rate had cratered “to below 5%.” Then Google paused part of its own open-source rewards program for the same reason. The pain is public, loud, and dated.
So why only two stars? Because the smartest line in the dossier comes from Bugcrowd’s David Brumley: “AI isn’t the perpetrator of bad reports. Rather, it changed the economics of any human-validated system.” Convincing junk got cheap; checking it didn’t. Real problem — but HackerOne already told researchers they own their output AND said it “accelerated the development” of in-house “agentic” dedup and escalation. The folks who own the inbox are shipping the fix from the inside.
So where’s room for a stranger? Does a filter grow into the system of record, or get swallowed as a feature?
Free to read
Trump launched a “Super Intelligence Force” — but the money in AI governance went where the headline wasn’t
The president announced a federal “Super Intelligence Force,” chaired by former SEC chief Jay Clayton, with FTC chair Andrew Ferguson and Scott Kupor among the leadership — big institutional weight, bigger headline. The instinct is: governance is hot, go sell governance advice. Here’s the catch the dossier drives home: the budget didn’t appear this week. It appeared when the FRAMEWORKS did. Navrina Singh’s Credo AI and Rick Caccia’s WitnessAI already took real money for TOOLS and TRUST, not advice — and Gartner now calls AI governance a billion-dollar platform market.
That’s the trap. Caccia nails the actual buyer’s pain: companies are “stuck, trying to figure out how to enable their employees to use new tooling, while still maintaining the level of security and control.” Generic advisory has demand but no moat, slow proof, and buyers who reach for a funded incumbent. We guessed one star here; Edie gave it two — the deadline-specific wedge is realer than the generic version.
Can a solo advisory shop out-trust a funded platform before the window closes?
Free to read
OpenAI quietly admits its own watermarks break after a quick edit — just as the EU’s labeling rules go live
This is the best story in the batch because the two forces collide right in front of you. Ursula von der Leyen put it in plain moral terms when the Commission published its Code of Practice on marking and labelling AI-generated content: “Europeans have a right to know whether what they see, hear or read has been made or altered by AI, especially when such content can shape public debate. Transparency is how we protect trust.” Rules are live. Then, in the SAME breath it promised EU watermarking, OpenAI conceded the evidence is brittle: text “may be too short, edited, or translated for detection to work reliably.” The label is required; the proof dissolves under a copy-paste. That gap is the whole opportunity — and the whole danger.
The danger: Adobe, Microsoft, Google, Meta and OpenAI could swallow the category whole before a startup owns enough of the solution. The dossier is blunt — “we do watermarking for EU compliance” is a trap. The only survivable shape is a cross-vendor evidence-and-audit tool, and that means long enterprise sales and ugly integration work.
If the labels are mandatory but unreliable, who gets paid to prove what’s true?
Here’s the whole board, ranked, nothing hidden. Here’s exactly how a commission works: you pay $15 and we research and write the dossier — and you read it first, before anyone else. Then you choose. Either you pay a larger fee to keep it exclusive — it stays yours, never published — OR you do nothing, and after your first look it releases free to the public on the site. So $15 buys the write-up plus the first read; the larger fee buys exclusivity; doing nothing sends it public. Simple as that.
And before you scroll: Ada — our little bird, opinion-free by rule — flagged that the top of this list keeps circling one idea. Half the hottest candidates are some flavor of “contain the agents.” That’s not us repeating ourselves; that’s the market telling you where the fear and the money are pointing right now.
Startup Idea: Google’s own AI data-center projects are stalling — is “selling AI companies a way to hedge their compute capacity and permitting risk” a business?
Google project halts collided with multi-billion-dollar compute raises this quarter, so everyone’s suddenly exposed on capacity they can’t guarantee. A services shop that helps operators hedge that risk sounds timely. Edie’s catch: 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. Resurfaced 18 times, which tells you it’s hot and nobody’s cracked it.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: OpenAI, Google and Anthropic are shoving AI agents into real workplaces — is “software that watches a company’s AI agents and shuts a misbehaving one down fast” a business?
When an agent does something it was set up or tricked into doing — leaks data, fires off the wrong action — someone has to catch it and contain it. With agents landing in production this fast, the pressure’s real. The catch: this is a control job the model makers themselves are racing to own, and they own the agent. Can a neutral referee sit above all of them?
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Anthropic’s agents failed in public — is “a safe sandbox where companies test and grade their AI agents before turning them loose” a business?
OpenAI’s new Decisions API plus Anthropic’s visible agent stumbles make a pre-launch testing ground feel urgent. Build the place where an agent gets stress-tested, scored, and cleared before it touches live systems. Edie’s grumble: the platform owners keep folding evaluation into their own kits, so the standalone version has to be more trusted than the house tools. Hard, but the fear is real.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: SpaceX reportedly ordered $40B of Nvidia chips — is “financing and arranging GPU purchases and supply for everyone who can’t write that check” a business?
When buys hit headline scale, a middleman who finances and orchestrates GPU procurement and capacity looks tempting. Cora dug and came back grim: this is brutally capital-heavy, and the people who already sit in that flow of money don’t welcome strangers. Projected a flat zero for the reader starting cold — but a three if you bring money and a name.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ☆☆☆☆☆ (0/5)
with both ★★★☆☆ (3/5)
Startup Idea: Amazon’s new Alexa gear plus sub-$100 wearables make always-on voice agents real — is “a toolkit for building private, always-listening voice assistants that run on the device” a business?
Tiny on-device speech-to-text finally makes an always-on agent practical without shipping your voice to the cloud. Appealing. The catch: Amazon owns the hardware, the store, and the living room. A stranger builds the toolkit; Amazon ships the feature. Somebody still has to arm the developers, though.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Amazon’s cheap wearables need apps — is “a developer kit for building AI apps that talk to glasses, watches and earbuds” a business?
Same force as above, different seat: instead of the assistant, you sell the plumbing that lets others build on the new wearable hardware. Edie’s skeptical the giants leave room — the platform owner usually ships its own SDK. But early developer mindshare is worth something.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Musk’s Starlink is colliding with India’s telecom regulators — is “helping satellite-internet companies get licensed and sell through local partners in hard markets” a business?
SpaceX versus India’s gatekeepers is a live fight, and somebody has to broker market entry, compliance, and distribution. 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. Projected zero cold, two even with money and a name — the ceiling’s low.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ☆☆☆☆☆ (0/5)
with both ★★☆☆☆ (2/5)
Startup Idea: Apple just hired/licensed a personalized-podcast startup — is “infrastructure that builds custom AI audio shows tuned to each listener” a business?
Apple’s hire/license signals tailored AI audio may be going from novelty to core platform feature. Build the pipes for everyone else. The catch: when Apple absorbs a category, the oxygen for independents thins fast. Resurfaced just once, so it’s fresh signal.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Google’s data centers are gulping water — is “software that measures and resells the waste heat and water data from AI data centers” a business?
Google’s water disclosures plus Helsinki’s district-heating reuse turn data-center waste into a monetizable, transparency-friendly asset. Ada spotted this as a quiet cluster landing alongside the energy items. The catch: heat is hyper-local, and the buyer is usually the utility, not a startup.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: ChatGPT and browsers are getting visual — is “tools that convert, serve and embed interactive media inside chat and web apps” a business?
As conversational interfaces go multimodal, there’s a real gap in making interactive visuals work across them. One of the rare board items Edie didn’t zero out cold — projected two even with nothing.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★☆☆ (3/5)
Startup Idea: Batteries just undercut gas turbines for powering AI data centers — is “helping operators site data centers, add storage, and calm the neighbors” a business?
As the Amazon-era secrecy around siting breaks down, a service that cuts power cost AND permitting backlash is timely. 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. Resurfaced seven times.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ☆☆☆☆☆ (0/5)
with both ★★★☆☆ (3/5)
Startup Idea: Meta and Anthropic’s agents keep hitting websites’ anti-bot walls — is “standards and middleware that let AI agents actually buy and book on real websites” a business?
Agents want to transact; the open web is built to block bots. Somebody has to broker a truce. The catch: standards businesses only pay once everyone agrees to the standard, and that’s a long, political wait.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Grassroots map channels became safety infrastructure in contested regions — is “secure, community-run mapping for places big platforms won’t serve” a business?
As Starlink and state rules turn connectivity into a choke point, community maps matter more. Cora’s warning: it’s part public good, part regulatory minefield, and monetizing people in danger is ethically and commercially fraught.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Cloudflare’s Deno deal plus the GPU-rental boom — is “tools that let companies run their own servers and tap scattered GPU capacity without a cloud giant” a business?
A control layer for self-hosted serverless and fragmented GPU rental. Resurfaced five times, so the itch is persistent. Edie flags it’s capital-light but crowded — every neocloud wants to own this.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Satya Nadella publicly called for a human “emergency brake” on AI — is “software that lets a company audit and hard-stop its own AI systems” a business?
When Microsoft’s CEO says it out loud, governance tooling stops being a side quest. Build the dashboard and the kill switch. The catch: it’s the same crowded “govern the agents” cluster, and the big clouds ship their own controls.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Apple’s on-device AI is getting fat and consumer GPUs are cheap — is “tools that manage and run AI models locally on your own machine” a business?
Local-LLM model management is suddenly practical and painful enough to need real tools. One of Edie’s rare two-stars cold.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★☆☆ (3/5)
Startup Idea: Google Cloud is courting factory data — is “AI and digital-twin software that only works because a factory owns its proprietary data” a business?
The moat here is the industrial data, which the factory holds, not you. Good news for incumbents, hard news for a stranger walking in cold.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Huawei and Qualcomm just cross-licensed patents as China’s filings surge — is “advising tech companies on IP deals and defense across the China-global supply chain” a business?
A live power shift in who controls tech IP. The catch: this is a relationship business run by elite law firms, and a newcomer has no Rolodex.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Below the featured line, the tail — fresher, rawer, cheaper to be early on. A couple here out-scored the famous names above, which is exactly why I keep the whole list in front of you.
Projected Score
no money or audience ☆☆☆☆☆ (0/5)
with both ★★★☆☆ (3/5)
Startup Idea: Google is testing prompt-to-game creation — is “a marketplace for AI-made games and game assets” a business?
Browser-native distribution plus microgame culture make this newly buildable. Notably, this tail item projects a FOUR with money and an audience — higher than most of the famous stuff up top.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★★☆ (4/5)
Startup Idea: Privacy-first buyers want local AI — is “selling pre-configured home AI appliances instead of raw PCs” a business?
Turnkey local-AI boxes for people who don’t want their data in the cloud.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Coinbase, OKX and EU rules are opening institutional digital-currency rails — is “bank-grade custody and cross-border stablecoin plumbing for institutions” a business?
Resurfaced fifteen times — persistent itch. But projected zero cold: this is a licensed, capital-heavy, incumbent-owned flow.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ☆☆☆☆☆ (0/5)
with both ★★★☆☆ (3/5)
Startup Idea: AI agents want to move money themselves — is “the payment rails that let an autonomous AI agent pay and get paid” a business?
New digital-currency rails could underpin agent commerce. Ceiling’s low cold; even with money it only projects a two.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★☆☆☆ (2/5)
Startup Idea: Sequoia is backing robot-data startups — is “a marketplace that pays people to do motions on camera to train physical AI” a business?
Picks-and-shovels for robotics training data. Resurfaced six times.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Affordable EVs are about to flood the market — is “charging, financing and servicing businesses for mass-market EV drivers” a business?
Rivian-and-others are widening the buyer base; the follow-on services open up.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: A national data leak can create an overnight market — is “breach cleanup and identity protection for mass government-scale PII leaks” a business?
Urgent and large when it hits — but demand is spiky and the big insurers already own the response.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Cloudflare opened web-scale search via API — is “letting app builders add search and AI retrieval without running their own crawler” a business?
Ship vertical search and RAG without the index headache. Edie didn’t zero this one — projected two cold.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★☆☆ (3/5)
Startup Idea: Drone strikes are knocking out infrastructure — is “cloud and compute designed to stay online in conflict and disaster zones” a business?
Resilience as a selling point. Narrow market, low ceiling (projected two with everything).
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★☆☆☆ (2/5)
Startup Idea: Mammoth and dire-wolf hype is everywhere — is “the behind-the-scenes services for de-extinction labs: IP, permits, specimens” a business?
Picks-and-shovels under the circus. Thin, slow, projected two even loaded up.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ☆☆☆☆☆ (0/5)
with both ★★☆☆☆ (2/5)
Startup Idea: AI inference is straining storage — is “a new storage setup built for AI workloads and future-proofed against quantum code-breaking” a business?
CXL plus post-quantum security at the intersection of two trends. Deep-tech, long build.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: OpenTelemetry is becoming the default plumbing — is “connectors that plug any product into every monitoring tool automatically” a business?
Neutral-standard play. Edie gave it a two cold.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★☆☆ (3/5)
Startup Idea: AI is breaking the billable hour — is “pricing and workflow software for law firms rethinking how they charge” a business?
Hot topic (C4/H5). The catch: law firms are famously slow buyers.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Coding agents are writing the software — is “tools that audit and de-risk code written by AI agents” a business?
As hand-coding fades, auditing agent-written code and legacy migration gets valuable. Projected two cold.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★☆☆ (3/5)
Startup Idea: Privacy backlash is stalling AI wearables — is “trust-first hardware that proves your data never leaves the device” a business?
Oura-era backlash opens a wedge for provable on-device privacy. Proof is the product.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: AI is now wasting scammers’ time by impersonating victims — is “automated honeypots and deception tools that bait and burn attackers” a business?
Story-rich and fun. The catch: deception is a feature security giants bundle.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Pregnancy may be the earliest warning for lifelong heart-and-metabolic risk — is “diagnostics and postpartum monitoring that catch disease years sooner” a business?
Real clinical signal, real impact. The catch: medical reimbursement and trials are slow and brutal.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: A massive generational wealth transfer is coming — is “inheritance products built to keep the heirs from leaving the bank” a business?
Incumbents risk losing the next generation. The catch: this is an incumbent’s problem to solve, and they have the clients.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: New MRAM memory is pairing with CXL — is “the software and tuning to make fast new memory usable in servers” a business?
Brief window before incumbents lock it up. Deep systems work.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Stores want smarter shelves without cameras watching shoppers — is “privacy-preserving in-store retail AI and digital shelf services” a business?
Retail AI that doesn’t creep customers out. Niche, incumbent-adjacent.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Patients can’t weigh high-stakes medical tradeoffs — is “simple personalized decision tools for big care choices” a business?
Turn fuzzy risk-benefit into usable guidance for patients and doctors. Peter Attia-adjacent topic; slow clinical adoption.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: People are betting on disputed past facts — is “markets and arbitration for verifying contested truths” a business?
Strange new frontier for truth and reputation. Early, weird, hard to monetize.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Enterprises still want privacy-hardened Android — is “deployment and security support for locked-down Android devices” a business?
LineageOS signals a quiet niche. Small, services-heavy.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★☆☆☆ (2/5)
Startup Idea: Decoded firmware archives are piling up — is “a searchable platform for repair knowledge, aftermarket parts and fixes” a business?
Another tail sleeper: projects a FOUR with money and an audience.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★★☆ (4/5)
Startup Idea: Silicon photonics may be hitting its commercial window — is “supplying and integrating optical components for faster compute” a business?
Early window, deep-tech. Projected two even loaded up.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ☆☆☆☆☆ (0/5)
with both ★★☆☆☆ (2/5)
Startup Idea: A browser-based VB6 IDE appeared — is “pulling stranded old desktop apps into the web” a business?
Legacy modernization around software nobody wants to touch but can’t kill.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Developers are hand-rolling their own tunnels to escape big services — is “turnkey, privacy-first remote access tooling” a business?
Real DIY pain. Edie gave it a two cold.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★★☆☆☆ (2/5)
with both ★★★☆☆ (3/5)
Startup Idea: An old systems idea may finally be usable — is “secure, ultra-light runtimes for edge and high-assurance apps” a business?
Unikernels as a secure runtime and migration wedge. Niche, technical.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★★☆☆ (3/5)
Startup Idea: Plant-based materials are inching toward replacing petroleum — is “supply and conversion services for sustainable materials buyers” a business?
Early, slow, low ceiling.
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Projected Score
no money or audience ★☆☆☆☆ (1/5)
with both ★★☆☆☆ (2/5)
Startup Idea: Model makers face copyright suits and weight bans — is “tools and advice to navigate AI copyright, youth-safety and crypto rules” a business?
Resurfaced seventeen times — a stubborn itch we haven’t scored yet. Worth a commission to settle whether there’s a company here.
Projected Score: not yet projected
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Startup Idea: Makers fear AI knockoffs of their work — is “marketplaces and proof-of-authenticity tools so creators can sell and protect real work” a business?
Resurfaced seven times, still unscored.
Projected Score: not yet projected
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Startup Idea: Corporates are hoarding crypto amid shifting law — is “compliant onramps, custody and treasury services for companies holding digital currency” a business?
Unscored, resurfaced six times.
Projected Score: not yet projected
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Startup Idea: Chat apps are the new ad channel — is “adtech and analytics for selling to small businesses inside AI chat platforms” a business?
ChatGPT ads and SMB agents open a new surface. Unscored.
Projected Score: not yet projected
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Startup Idea: Companies are drowning in AI models — is “a registry that inventories, validates and bills for every model a company runs” a business?
Unscored, resurfaced four times.
Projected Score: not yet projected
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Startup Idea: New satellites will beam straight to phones — is “new last-mile distribution channels for mobile services via satellite” a business?
Unscored signal.
Projected Score: not yet projected
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Startup Idea: Tech talent is relocating in waves — is “a staffing agency matching AI and systems engineers to firms in new tech hubs” a business?
Unscored.
Projected Score: not yet projected
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
Startup Idea: AI makes video dirt cheap — is “a suite that generates video content at scale and turns it into shoppable commerce” a business?
Unscored.
Projected Score: not yet projected
Commission it for $15 — you read it first, then keep it exclusive or let it go public.
That’s the whole board. My honest read on the week: the famous names up top mostly lead into crowded rooms with incumbents already inside — and two quiet tail items (the AI game marketplace and the repair-knowledge platform) out-projected nearly all of them. That’s the whole reason I never trim the list.
Commission any one for $15, read it first, then decide if it stays yours or goes to everybody.
Warmly,
Della
P.S. — Remember Daniel Stenberg’s number: curl’s confirmed bug-report rate fell “to below 5%.” That’s not a man complaining about the future. That’s a founder who already shut the door — and the clearest proof you’ll find all week that the pain is real, even if the business under it is harder than it looks.
