cole

Shipped this yesterday. Reply to this tweet and I’ll tell you what our AI agent says your vibe is and what kind of business you should run.👇

moved here a couple weeks ago and it has beyond exceeded my expectations i did not realize how beautiful sf actually was until i saw it in person the people here aren't perfect, but i feel like that has turned into a bit of an echo chamber i think we can all agree

Our Al just told me i have "streetwear taste and supercar energy" and that i should start a sauna hat company. Shipped a new Locus Founder feature today. It researches your X profile and gives you a business idea to start.

@MertCologlu I haven’t yet! I’ll check it out

@coledermo Did you try Orgo ?

4 minutes 42 seconds. that's how long it took our AI to research me. Uploaded some photos, gave it my instagram, and the Locus figured out my whole vibe on its own. Matched me to supercar culture + music, built the brand, generated the site. You can now literally speedrun building an internet business.

@kleffew94 🫡

Cool x402 use case

when im a public company ceo i pledge to post just like i do now

@growing_daniel Blockchain layer 2 but make it ai

The real crime at Phia was stealing Honey’s playbook. Stealing a fellow startup’s fraud idea is deeply unethical

this is if you’re solo operations, i think if you have a team of people doing it’s lighter. generally just at any given moments ops has more moving pieces than eng, plus the amount of context you need to maintain is pretty brutal even with ai. it’s just a lot to keep in your head 24/7 esp while still trying to find time to focus on product and eng. being able to do it though is also probably one of the more distinct advantages you can have atm.

i thought engineering was hard, try operations 💀💀 you wake up and have new things every day, and you have to jump between them like a crazy person

@levelsio We’ve had like 10 separate german users hit us with exactly this lmao, however they do give the most granular feedback so can occasionally be useful i guess(?)

2 hours after a German signs up to your site

@heysamir_ @TownAI @Superhuman @Claude It’s funny how uniquely indefensible the crm market is for startups, I feel like people switch so frequently

*Running list of every app I no longer use b/c of AI super assistants (I use @TownAI)*: - @Superhuman: Moved back to free Gmail because Town AI has an in-app composer that pre-drafts messages in your style both on command and proactively (with your approval) - @Claude Design: Town makes great looking blog post banners, draft logo/wordmarks and mockups - Voice Memos: Town has built-in dictation with perfect transcription and it immediately becomes a task that you can then take action on - @attio: Moved to @twentycrm which is more extensible for cheaper since Town offers full AI analysis and structuring within Twenty - @calcom: Downgraded to free plan since Town can review my calendar and sender's and provide natural email follow-ups to figure out the right time - @typefully: Town keeps track of best times to send tweets, edits them and helps me do quick research - @lexdotpage: Town edits my blog drafts and helps me manage tone. Builds full formatted docs with charts and tables in Google Docs ready to publish - Apple Reminders: Migrated entirely to Town by screenshotting reminders and Town offered to help with at least 75% of them *Single-player utilities I'll build in-house to optimize costs given simple use case*: - Contracting/eSign: Great OSS out there already and this will allow me to host embedded within my website - Code Editor: Forking @boltdotnew to build opionated vibe code editor that lets me use GLM coding at fraction of @Replit - Meetings: Might use Anarlog, OSS version of Granola to build something more customized to my stack *Apps I still use given unique value moat*: - Physical Mail: @useStable: No other way to get business mail scanned and documented via API/MCP - Accounting: @puzzlefin: The best AI-native accounting ERP on the planet and something I never want to build on my own given complexity - Business Phone @gowithquo: No other way to get API-enabled business phone - Payments @stripe: Best way to have payments in apps and in contracting software - Banking @rhobusiness & Waldo: AI-native and API-driven workflows to support banking & treasury - Domains: @spaceship: Best combo of value, simplicity and API access x.com/heysamir_/stat…

one of the things i value most about the company and culture i’ve built to date is that we’re all friends. it sounds silly but if you need to maintain working harder than everyone else, the only way you’ll maintain sanity is by doing it with talented people who you love spending time with. i need everyone to love it here since otherwise the pain of maintaining becomes far too high. so now the question becomes how do you create that. here’s what i’ve found works best for us: 1. always prioritize hiring in-network whenever possible: if you have an existing relationship with someone talented, they’re likely already convicted in you and in some minor way culturally aligned. a key caveat here is to also avoid group think and structure team ideation and decision making in a way that inherently prevents this. this is also the obvious risk of hiring friends. 2. filter during interviews: ask dumb questions sometimes that seem completely irrelevant to see how people reply. for us it’s the “billion lions vs. every pokemon question” but you need to pick something that you know your team would vibe with. this is a really easy way to get a cultural read of someone. 3. work trial: consistent learning we’ve had is that 1 week is not long enough. after a week you just start to see someone execute typically, if you want a real read of quality, 2 weeks is the bare minimum. put it this way: a week gives you a read of how someone pushes for an offer, 2 gives you a read of their execution, problem solving, team fit, and more. i obviously did not touch at all here on hiring talented people, that is a separate problem and is the reason why work trials and good heuristic filtering and assessments are important, but i believe culture to be near equal. if you hire someone 90% as talented but 100% culturally aligned, they’re likely willing to put in 10-30 more hours weekly. this additional time usually yields far more marginal benefit than the increase in talent (and takes basic arithmetic to see why). these are some rules but there i’m sure are many more. curious on others thoughts on this matter and what they’ve found works. other founders, what are your thoughts on this? what works and what doesn’t? what kind of importance do you put on this type of alignment?

i want to get back into fashion when i have the chance i feel like my drip has fallen off immensely (4th pic is current) maybe post ipo

we rebuilt our onboarding 3 times in 3 months first version: too many screens second version: too few third version: hopefully right the embarrassing part is we shipped v1 to paying customers. they built businesses on a flow we didn't know was broken (but learned quick) they didn't complain. they just used it anyway. ship fast and don't over-optimize for perfection. you only have data once it's in someones hands and that's the only way to make the product better

no it’s because average people are broke. the data supports this and sits entirely contrary to this tweet. this feels very out of touch.

our ai has written 45,273 chat messages to users humans on our team have written 10,640 the company is mostly run by the thing we sell and honestly i didn't plan it that way. it just turned out the tool we built for other people was the tool we needed most ourselves 56,133 messages total. roughly 4 out of 5 were written by the product

Excited to be building at the forefront of payment infrastructure for agents. Looking forward to sharing what we’ve been learning with merchants and builders at @paywithlocus

@sxnatra_ i don’t think so considering the frequency, for some reason on market research the weather data has brought better results? it’s quite strange. transparently not always tokyo but i guess it uses it to determine more granularly types of clothes (as an example) that’ll sell

at some point our ai bought weather data 15,643 times i have no idea why. we don't use weather data. it's not in any prompt. somewhere in the agent's decision tree it decided knowing the temperature in tokyo was critical to starting businesses we left it running. it's probably bought more by now

@ArshJain08 Thank you sir🫡

@coledermo most goated hoodies ever

thinking about doing a giveaway of one of our hoodies as a prize for whoever can build the best business on Locus Founder after a week. dm me for a code for 50% off and we’ll help you get started, next friday whoever can get the furthest I’ll ship them a free Locus hoodie