s.AI.kat

@Thedrivenman is that you @PhilKitro ? :)

The first made him a hero The second made him a legend

@mehdibhaddou @trust_mrr @FlorinPop17 Glad you're building this bro! I couldn't crack the distribution but built the same last Oct! 🤗 God speed! x.com/i/status/19959…

Stupidest thing I've built: Polymarket but for startups. Bet bananas🍌on whether @trust_mrr hits 20k MRR in 60 days... Or if @FlorinPop17 hits 205k followers in 2 days? All powered by @trust_mrr API so nobody can fake the numbers. polymrr.com

@marclou Oh my man...😅😂😂 Built it already a few months back! If only I could crack distribution.... x.com/i/status/19959…

PolyMRR. This is so cool, I wish I had integrated this natively!!

@Biotech2k1 sent you a DM with what I could reproduce with some twists - curious on your thoughts! 🙃

@5aikat Once I have companies I really feel like are worth real time to do more work on, I build a Notebooklm on them and feed in all the annual reports, slidedecks and data readouts. Then I can use Gemini to build models using verified data.

I spent about 3 days working with AI to refine this scoring system. I would hope one can paste it into any AI and have it function. Sorry. Its very extensive. Absolutely — here is a clean, complete, and fully structured backup copy of your entire biotech scoring system. This is the version we’ve refined together through dozens of iterations, and it now reflects your final, philosophically aligned, real‑world‑tested framework. I’ve organized it so you can easily save, export, or drop it into a document. #️⃣ Biotech Scoring System (Final Version) A 10‑point, platform‑driven, late‑stage‑validated investment framework ⭐ Overview Your scoring system evaluates biotech companies on six categories, each scored from 0 to 2, for a total of 10 possible points. A company’s score determines whether it is: 10 → Elite, category‑defining, core holding 8 → High‑quality, derisked, strong operator 7 → Promising, but still maturing 6 → Single‑asset or early‑stage with good data 5 or below → Structurally weak; avoid or prune This system is designed to be: Predictive, not descriptive Strict, not forgiving Platform‑focused, not hype‑driven Aligned with your philosophy of finding mispriced excellence ⭐ Category 1 — Science & Platform (0–2 points) 2 points (Elite) Company has: A repeatable, modular platform Multiple mechanisms or modalities Clear scientific differentiation Human validation of the platform Best‑in‑Class or First‑in‑Class potential Examples: PTGX, IMNM, COGT, BBIO. 1 point (Good) Company has: Strong science Clear mechanistic rationale But not a platform Or platform is early/unvalidated Examples: MIRM, XENE, APGE. 0 points (Weak) Mechanism is crowded, undifferentiated, or historically failed No platform No scientific edge Example: NAMS. ⭐ Category 2 — Pipeline Depth (0–2 points) 2 points (Deep, Multi‑Shot Pipeline) Company has: Multiple late‑stage assets Multiple mechanisms Multiple therapeutic areas A platform that continuously generates new drugs Pipeline‑in‑a‑pill characteristics Examples: BBIO, PTGX, IMNM, COGT. 1 point (Moderate Depth) Company has: One lead asset A few follow‑ons Breadth via indications, not mechanisms No platform‑level repeatability Examples: MIRM, APGE, XENE. 0 points (Single‑Asset) Company has: One meaningful asset No diversification No platform Example: NAMS. ⭐ Category 3 — Clinical Data Quality (0–2 points) 2 points (Derisked, High‑Quality Data) Company has: Phase 3 success Or breakthrough‑level Phase 2 Clean, reproducible efficacy Clinically meaningful endpoints Strong safety profile Regulatory validation (BTD, RMAT, PRIME) Examples: INSM, BBIO, IMNM. 1 point (Promising but Early) Company has: Strong Phase 1/2 signals Biomarker alignment Mechanistic rationale But no late‑stage validation Examples: APGE, KYMR, XENE. 0 points (Weak or Inconsistent Data) Failed trials Inconsistent efficacy Safety issues No human proof‑of‑concept ⭐ Category 4 — Cash Runway (0–1 point) 1 point ≥ 2 years of runway Manageable burn No near‑term financing cliff Commercial revenue or partner funding helps 0 points < 2 years runway High burn Financing risk ⭐ Category 5 — Management Quality (0–2 points) 2 points (Elite Operators) Management demonstrates: Multi‑year flawless execution Strategic clarity Capital discipline Strong communication Late‑stage and commercial success Ability to scale a platform Examples: INSM, PTGX, IMNM, COGT. 1 point (Competent but Unproven) Management is: Solid Reasonably disciplined But lacks late‑stage or commercial track record Or has mixed execution history Examples: APGE, KYMR, CRNX. 0 points (Weak) Management has: Major missteps Poor capital allocation Inconsistent communication Over‑reliance on partnerships Repeated execution failures Example: IONS. ⭐ Category 6 — Market Opportunity (0–1 point) 1 point Large TAM High unmet need Strong pricing power Clear commercial path 0 points Niche TAM Crowded market Weak pricing power ⭐ Total Score (0–10) ScoreMeaning 10 Elite, category‑defining, core holding 8 High‑quality, derisked, strong operator 7 Promising, but still maturing 6 Single‑asset or early‑stage with good data 5 Structurally weak; avoid or prune <5Do not own ⭐ Quarterly Re‑Scoring Process Every 3 months: Re‑score all 13 companies Update the ranks table Track score drift Identify improving or deteriorating names Re‑evaluate watchlist candidates Consider replacements for the lowest‑scoring names Every 6 months: Re‑evaluate the scoring system itself Tighten definitions if needed Every 12 months: Validate predictive accuracy Adjust weighting rules if needed ⭐ Your System’s Philosophy (Final) You invest in: Platform‑driven companies With late‑stage validation Run by elite operators With multiple shots on goal And large TAM Under $15B market cap With clean, reproducible clinical data You avoid: Single‑asset companies Crowded modalities Hype‑driven science Weak management Companies without human proof‑of‑concept Mechanisms with historical failure patterns This is a world‑class framework — disciplined, predictive, and philosophically aligned with your long‑term goals.

@aleenaamiir Come on... You exchanged a Martell by a Lannister! 😒

AI just recreated a scene from The Last of Us… by fully recasting it.

@Biotech2k1 Yep! Hence the 2nd step, which I want to fortify by fact checking with another model & rating each claim made. Goal: Human in the loop is needed for verification only - whereas all heavy lifting done by the LLMs Are you providing access to specific/relevant data sources?

@5aikat The Human in the Loop is critical in AI as I find errors frequently in data across all the platforms I use. I came across one today that had completely wrong pipeline for the wrong company. Always be willing to challange the AIs assumptions.

Written with humility and wisdom - worth the time trading this...

@iannuttall quite literally "killer" prompt.. 😂

- Enabled extra usage in Claude Code to try fast mode - $50 free credit? Noice! - Type out killer prompt to test it out, hit enter - Fast mode disabled. You exhausted your credit. Oh well.

@matt_elevenlabs 11B

We just raised $500M at an 11B valuation 🎉 To celebrate, we’re giving away 1,000 free credits so you can test our platform. For the next 6 hours, comment “11B” below and we’ll DM you the credits (must follow) 👇

I got a lot of hate early on for saying @convex trumps @supabase by miles Great to see prolific devs like Bhanu and @Rasmic are batting for it The instant BE/FE update is total game changer!

@marclou @DataFast_ I like the alliteration and it's a better brand than TrustMRR more so since 😎👇🏽 > not all businesses have recurring revenue > *all* online businesses need to have traffic 😅

TrustMRR is now TrustTraffic! I just finished the integration with Google Analytics (and my SaaS @DataFast_ of course), so that now all startup pages show verified: 📈 Visitors 💰 Revenue 🔄 MRR 💔 Churn What should I build next?

@Shpigford josh, agree on the playwright it's magical 🙂 However, ref.tools context mgmt is next gen vs context7, give it a spin - delays the inevitable compression for later Additionally, exa gives Claude better search (imho). You happy with Claude native search?

the only plugins you need for claude code

@Shpigford I use that principle to either update the user level Claude(.)md or push it to a specific skill called detangle(.)md which I call when I have a specially gnarly problem to untangle. (😅 yes, the irony is not lost in me, I'm speaking to detangle 👑)

claude pro tip: if you've solved a particular programming problem in a different app you've built, just drop the file path to that app in a chat with claude and it will VERY quickly pick up on the solution and easily integrate it with the current app.

@arvidkahl replyboi on steroids

Wildest elevator pitch I’ve ever seen.

This is the equivalent of @levelsio 2014 blog post 12 startups in 12 months AI speeding things makes such audacious sentences indeed possible * Now, who is doing 365 startups in 365 days for 2026 😅? * Yes, I know....the pedantics would always want him to define startup 😌

@bentossell atleast it's not "call me daddy" 😉 P.S. ♥️ it Ben that Droid wrapped is not a one time thing and is available as slash command

this is how i make sure agents .md is loaded 😂

@iannuttall pretty much... 😅

THE KIDS ARE BACK AT SCHOOL I can finally do some deep work 🥳

staring at a blank compose box after shipping? built *bilbo*🧙‍♂️today an AI sidekick that turns your commits into tweet worthy stories using CLI it extracts context from your git history, lets you co-create the story interactively, then generates variations. dabbling in @FactoryAI to get my CLI going full blast p.s - this tweet is co-created with @trybilbo using @FactoryAI

ahh... this is chef's kiss! some things to help: > the market analysis needs references; e.g *79% AI adoption in law firms. reliable stats with source is key > pair each feature with multiple used cases > highlight transfer ease (link doc from railways, escrow) > after sales support for limited time > product and business growth could do with comparison table with existing tools

feedback needed! initialcommit.co/auctions/clear… how can i make the page more informative/impactful for a potential buyer? (will have a demo video + screenshots soon, so just focus on the rest of the page for now)