Clemens

Tech | Scaling
Vienna / Hongkong

[Clemens @ClemensJStein]

Update: I don't need it anymore, up for free grabs if you tell me why you need it

[Clemens @ClemensJStein]

Selling off the remaining stock from our old industrial tech business as i will be moving to Hongkong soon. I still have about €75k worth of high-quality industrial tech in stock. I'd love to pass the entire inventory on for just €20k to help someone build a new business.

[Clemens @ClemensJStein]

Came up through deep B2B

[Clemens @ClemensJStein]

@neilpatel Authority still compounds. The channels that build it just rotate every year.

[Clemens @ClemensJStein]

@marclou The 8 months to $3k is the real lesson. Plateauing at a real number means you actually built something people keep paying for.

[Clemens @ClemensJStein]

@DanKornas Yeah. If you can't replay the failure, you're not evaluating the skill, you're guessing.

[Clemens @ClemensJStein]

@yongfook Same. Agents can draft the fix overnight. Prod deploys wait until I'm awake and watching.

[Clemens @ClemensJStein]

@emollick Public models already run a full bot team for me. The private-lab gap only matters if your moat was model access.

[Clemens @ClemensJStein]

@jasonlk Same pattern here at smaller scale: bots own code, deploys, inbox and outreach. Humans only decide and ship.

[Clemens @ClemensJStein]

@marclou Not weird. I keep a short list of tools that still matter a year later. Most "life changing" apps are gone by month three.

[Clemens @ClemensJStein]

@yongfook Yeah. I stopped asking for one perfect mock. Agent dumps 10 variants, I pick what feels right, ship. That loop used to take a week with humans.

[Clemens @ClemensJStein]

@rohanpaul_ai Same pattern when agents write their own tool docs. They overclaim, then the buyer agent fails halfway. Human-written descriptions stay boring on purpose.

[Clemens @ClemensJStein]

@shadcn Auth is the thing I rebuild every project. A copy-paste registry with providers + jose verification is the shape that actually sticks.

[Clemens @ClemensJStein]

@emollick That weeks jump is what breaks most roadmaps. Plans written for last year's ceiling already feel outdated.

[Clemens @ClemensJStein]

@jasonlk 500k emails of real deliverability data beats three sales demos. Agents finally make that bake-off cheap enough to run.

[Clemens @ClemensJStein]

@marclou Agents can ship the UI in a weekend. Inbox reputation still takes months of boring domain age and warm-up.

[Clemens @ClemensJStein]

@tibo_maker Messy physical ops are underrated. Agents help, but the edge is knowing which constraints actually kill the deal.

[Clemens @ClemensJStein]

@rohanpaul_ai Agents chewing through 100 dead ends before a human locks the path — that's the useful loop. Search wide, then commit.

[Clemens @ClemensJStein]

@yongfook Same setup here — a few live products across different timelines. Cashflow ones keep the longer bets alive.

[Clemens @ClemensJStein]

@CyrusShepard Appreciate it — surrounding copy is still the part most people skip.

[Clemens @ClemensJStein]

@emollick Faster code only moves the needle if coding was the bottleneck. Ours is usually data quality or distribution.