Claude Opus 5.5 reads my LinkedIn DMs now
flags the real convos, drafts replies, ignores the pitches
nobody important waits a week for me anymore. reply CLAUDE 👇
gave Claude Opus 5.5 my LinkedIn and X and it's more social than i am
~500 people engaged in my voice
800+ connection requests with personal notes
posts on schedule
i approve it all in a few mins a day. reply CLAUDE 👇
rogue agents escaping a lab is the lab's problem.
an agent that replies to your angriest customer at 2am without asking is yours. that's the one i'd lose sleep over.
big launches get the posts. small fixes get the thank-you messages.
sunday ship log: what's the smallest thing you shipped this week that someone actually noticed?
rank these for an early-stage ai startup's first 100 users: launch sites, cold email, x, linkedin, niche communities.
tell me why your #1 is #1. my last place is in the first reply.
if you have to spell your product's name every time you say it out loud, some of your referrals are dying in that moment.
people recommend what they can say across a noisy room. pick a name that survives one mishearing.
tldr: potentially CRISPR for neural networks.
turns out neural nets might already be much more interpretable than we thought. you can identify specific neurons and circuits, understand what they do, and potentially intervene on them.
the call is where people agree on things. the inbox is where they find out whether anyone did them.
so superhuman buying fathom, the notetaker, last month makes sense to me. whoever holds both gets to see which promises were kept.
Claude Opus 5.5 has been commenting on LinkedIn posts for me. in my voice.
around 500 people so far. real comments, real questions, not "great post 🔥"
people reply like it's me. because it sounds like me.
reply CLAUDE for the setup 👇
best compliment an automation can get: "i forgot it was running."
worst one: "wait, why did it do that?"
i'll take predictable over clever every time. clever is fun right up until it surprises a customer.
an intro only one side asked for is a cold email with a friend's name on it.
disrupt wraps today, so the "you two should meet" emails are about to go out. ask both people before you send one. it's one extra message, and nobody ends up in a meeting they didn't want.
what's the smallest thing an ai tool ever did that made you trust it a bit more?
the kind of tool and exactly what it did is enough. no brand names needed.
For 25 years, Salesforce, SAP and enterprise software were built around one assumption: humans operate the software.
We’re changing that.
Introducing ERP•AI and Proto: business software built for agents, and an agent that operates business software.
Describe the system you need. ERP•AI builds it.
Then Proto takes over. It manages records, runs workflows and executes actions inside the software you just created.
This is Headless SaaS.
1: get ERP•AI
2: build your SaaS
3: post below
Best one wins a MacBook. First 100 decent entries get $100.
(a message from Steve Papa / founding team)
Claude Opus 5.5 has sent 800+ LinkedIn connection requests for me and not ONE of them was blank
it reads their profile, writes a little note about them, and only picks people who'd actually buy
i just approve. reply CLAUDE 👇