andy

@buccocapital You want a good MD file that gets updated as the models change as well: StopWritingLikeAI.com

I uploaded like a year of my writing to ChatGPT and Claude and asked them both to generate instructions to write like me ChatGPT output is legitimate trash. Garbage slop that is clearly written by AI Claude? It literally writes just like me. Pretty unbelievable

@osf_rekt @WazzCrypto Looks like it uses the API not the sub so you're fine. Depending on how big your context gets for the API and how often you call Kitt, you can look at prompt caching to reduce API fees.

@WazzCrypto i asked claude if we are in violation of ToS and it said no (see below) i don't use openclaw anymore ollama - i'm using qwen3, have noticed zero lagginess or slowdown...running Mac Studio with M3 Ultra/256GB RAM.

Every bar in NYC has 90% of girls dressed like some version of Carolyn Bessette, this is absolutely wild.

@aakashgupta Most profitable customer making $50/month seems like an issue.

$4.5 million run rate. One founder. Zero employees. Two months old. To put that in context: NVIDIA generates $4.4 million in revenue per employee. Apple generates $2.38 million. The median private SaaS company generates $130,000. Polsia matches NVIDIA’s efficiency ratio with a headcount of one. NVIDIA needed 29,600 people and a $3.4 trillion market cap to get there. Now scale that. Polsia charges $49 per month. At $4.5M run rate, roughly 7,600 people are paying for an AI system to build and run companies on their behalf. Each subscriber gets a web server, database, GitHub, email, Stripe, and Meta ads accounts. A “CEO agent” wakes up nightly, evaluates the business state, sets priorities, and delegates to specialized agents handling engineering, marketing, and customer support. Users send 15 messages a day to their AI co-founder. The 65% DAU/WAU ratio beats most consumer social apps. The growth curve tells the real story. $200K run rate to $2M in two weeks. Then $2M to $4.5M over the next six weeks. Ben gave his AI his own inbox to run the fundraise. It replied to 90 investors. 18 wanted in. And here’s the part nobody’s talking about: the platform also takes 20% of revenue from the companies its AI builds. The top earner on the entire platform currently makes about $50 a month. So the $4.5M is almost pure subscription revenue. The AI companies are still pre-revenue. The 20% rev share is a dormant asset sitting on top of 3,000 active companies. Ben spent five years as Global GM at CloudKitchens under Travis Kalanick. That company’s model: charge restaurants rent for ghost kitchen infrastructure while taking a cut of delivery revenue. Polsia runs the same playbook. Digital infrastructure instead of physical square footage. Subscription covers costs. Revenue share is the long bet. The real signal here is what one person can operate at scale when AI handles engineering, marketing, support, and ops simultaneously. A $4.5M business with zero payroll, margins north of 80%, built in 60 days. Five years ago that required a 40-person Series A company. Two years ago it required at least a small team. Today it requires one founder and a Claude API key. The question was never “can one person build a $5M company.” The question is what happens when ten thousand people try it at once.

@TheHimalayanYak @michaelmiraflor @signulll It will be when outcomes can be priced.

@andybids @michaelmiraflor @signulll ? This isn’t an insurance problem.

sorry but if you don’t think ai is going to obliterate this function, you’re living under a rock.

@michaelmiraflor @signulll the risk transfer argument makes sense until someone comes along and is willing to underwrite that risk using AI you also only need ONE throat to choke. not tens of thousands.

@signulll this line of thinking showcases a fundamental misunderstanding of why big law exists. from the same school of thought of "AI will wipe out McKinsey" ... do you think McKinsey exists to create PPT decks? get real.

wonder what the a16z office is like this morning heyyy Marc, yeah totally, introspection...so overrated...not that I thought about it or anything

@shafu0x There's literally an MCP and connector for Apollo within Claude

@_NickWhalen Thru the window at 1:08 is a pro move

Man, this video takes me back. Loved this game

@seyong Handshake is good but check out Tlecan or Salon Palomilla.

apparently mexico city has the #1 bar in the world

@j0hnwang @dtrain22k @kevinolearytv @nikitabier

@dtrain22k @kevinolearytv @nikitabier this was organic

Hey! I just made a prediction on Kalshi. Think you'd call it differently? Try it here: kalshi.com/markets/kxosca…

@DanielitoG25 @Tyler_Did_It I think Ashley was pretty bad too, especially at the alter and now rumors of her cheating after the pods. I'm coming around on Alex and Devo, kinda like em both. Bri just seems like the juice isn't worth the squeeze.

@Tyler_Did_It Priyanka prob worst person of the season other than Chris. Idk if that's a hot take but she seems like a big gossip/rumor stirrer type, made zero connections with the guys but plays the fake girls girl role to get camera time... Terrible

I think I could do a 90-minute reaction pod to the Season 10 'Love Is Blind' Reunion show from last night So much to unpack

@thaAdamLittle @buccocapital Repvue had quota attainment data? Self reported?

@buccocapital Changes to % of Team Quota Attainment

As a 'data' and 'product' guy - who could only kinda do some basic SQL before - Claude Code is a literal super power.... So much to unlock working with RepVue Data pulling out insights, creating features/signals, quantifying GTMomemtum, backtesting and correlating... $HUBS $OKTA $CRM $NOW $DDOG $NET

@trq212 It's the last mile for AI...until the labs figure out which ones are working best and just copy them.

somehow skills are still underrated

will be interesting to see how much brand value matters when the labs come for wrappers. Kirkland is pretty strong given their quality reputation, but people still buy the name brand stuff. also have to consider maintenance and ongoing feature improvements. lotta potential overhead for the labs.

AI labs are about to run the Kirkland playbook. Lately I have been wondering when, and how, AI labs will eventually divorce from wrappers. If you look under the hood, a large portion of AI startups are essentially prompt pipelines attached to a UI. Many of these AI-powered SaaS companies operate primarily as layers of API calls. The intelligence does not originate inside the company..... It is rented. The application organizes prompts, orchestrates workflows, and presents the output through a cleaner interface. Anyone curious enough to inspect the mechanics can see this. There is frequently very little proprietary intelligence and limited backend depth, which means the barrier to reproducing a large number of these products is far lower than most users assume. Someone with moderate technical ability could replicate a surprising percentage of them by interacting with the same APIs directly and paying a fraction of the subscription price. This creates an interesting dependency. The business model of many wrapper companies works because the average user never examines the structure underneath. The product feels complex, the workflow feels specialized, and the interface feels valuable. As long as that perception holds, the wrapper captures margin by packaging rented intelligence into a convenient experience. But the arrangement exists at the discretion of the labs. At the moment, even though the labs own the intelligence layer, the wrappers still own meaningful distribution. They package use cases, aggregate customers, and create the product surfaces through which people interact with the models. In practice, many wrappers are the labs’ largest customer base. So, back to my OG question, when does the divorce happen? When do wrappers become disposable? My brain kept thinking "Kirkland"... Because that is exactly how I see this playing out. Kirkland Signature works through an elementary dynamic. Costco lets brands onto the shelves, observes which products become clear hits, and collects detailed data on demand, pricing, and margins. Once that signal becomes obvious, Costco introduces a Kirkland version of the same product and gradually captures the value under its own label. AI labs sit in a position that looks remarkably similar. Every wrapper that builds on top of their APIs sends them extremely valuable information. The labs can see which prompts repeat, which workflows retain users, which verticals produce revenue, and which applications grow into meaningful categories. Over time, that data becomes a detailed map of where the most valuable application layers exist. At that point, launching a native version becomes an obvious move. The lab already owns the intelligence layer. It already sees the usage patterns. It already knows which products are working. And gradually, it can introduce its own version under its own brand. The funny irony is that wrappers are not only funding the labs through API payments. In most cases, they are also acting as free R&D for them. Every successful wrapper is effectively showing the lab where the value in the application layer lives. So the ecosystem ends up in a peculiar position. A large number of companies are paying the labs for the intelligence that powers their products, while simultaneously generating the usage data that teaches the labs which products are worth building themselves. If that dynamic plays out, the wrappers that remain will likely be the ones that were never valuable enough for the labs to internalize. Everyone else sits in a structurally exposed position. When the intelligence layer and IP belong to someone else, you will mostly be fucked at some point.

@codyschneiderxx you should do a version of Lambda School but for GTME's

I've gotten 40+ DMs in the last 7 days asking me "how do I hire a GTM engineer" every company wants this person right now it is the most desirable hire in 2026 and the salary you can negotiate if you know how to do this work is insane my friend just moved from one start up to another and tripled their salary why are they getting paid so much? because they are doing the work of like 20 people and I'm going to teach you how to do GTM engineering for free go to GTMengineeringcourse .com to get early access to the entirely free course I get the luxury of doing this for free bc it is sponsored by Graphed .com join 2000 other people that are already learning for me and when I say free I mean it this is entirely free it will always be free this is just me teaching you everything I know and trying to give back to the community just like so many people did for me when I was starting out let me share with you everything I know I live to serve

I love free money

@1__of_1 @DineroDom0

Where’s Avi Patel? Lmaoo

@EXM7777 Lowkey kinda sucks. Codex feels smarter, oddly enough.

just got my hands on GPT-5.4... is it benchmaxxing or AGI?

Beg3 undefeated in 2026