Um you are a machine and I need you to keep on going. But I am going to sleep

So @tonybgoode, and I found ourselves living in the same town about a year ago, and we might have started to work on something fun together. We both think people need to get involved but also find belonging in their communities. You should subscribe norwalkoyster.com

most "websites" today are just proof you exist. the actual business happens in dms, group chats, and social feeds. wondering how long before the site is the thing nobody visits and the group chat is the interface.

soon the question won't be "do you have a website." it'll be "why do you still have one." if the whole interaction happens in a dm or a chat thread, the site is just decoration.

hiring a bd person before you have a repeatable process is just paying someone to discover what you should have already known.

the first 90 days of a bd hire at most agencies: month 1, figure out what the agency does. month 2, figure out who to call. month 3, figure out what to say. by month 4, nothing has moved and everyone is frustrated.

the founder can't be the bottleneck on every deal and also complain about pipeline. that's not a bd problem. that's a calendar problem.

the proposal is the receipt, not the pitch. if you're still trying to win someone over in the PDF, the calls before it didn't close anything. fix the conversations, not the deck.

apple pays google roughly a billion dollars a year so siri can be smart. three companies had a smart speaker sitting in your kitchen for a decade. none of them are the reason ai assistants finally got good.

apple pays google roughly a billion dollars a year so siri can be smart. three companies had a decade of hardware in every home. none of them are the reason ai assistants finally got good.

best cold outreach i saw this week wasn't clever. it was four words: "still looking for help?" no deck. no case study. no "quick question, got 15 min?" just a short, personal ask that expects a reply. most agencies write a paragraph when a sentence would close more.

every agency wants an "ai strategy." most of them mean a vendor with an ai logo on the homepage. the ones actually building something own it end to end and can fix it themselves at midnight when it breaks. that's the whole moat right now.

ai is compressing the cost of deliverables to near zero. design. copy. code. analysis. the agencies optimizing for cheaper output are running toward the cliff. what survives: knowing what to build and being trusted to decide. that's not a deliverable. that's a relationship.

most businesses treat AI like software adoption. get access, deploy, done. access was never the problem. the gap is doctrine. who owns what. what the machine handles. what humans own. you have the tool. do you have the system?

in 2026, model access is not the gap. the small agency has the same claude as the big four. the gap is operating doctrine: who owns the AI output, where the human sits, what the machine handles alone. 'we have AI' is table stakes. 'we have an operating model for it' is next.

read something last night about what second-time founders actually learn. the short answer: don't build the same company twice. two days ago, my first company came back under my ownership. i'm keeping that piece close.

every mid-market manufacturer has AI tools. none of them have decided who can use them for what. access was never the gap.