Jake Deichmann

[Jake Deichmann @jakedeich]

Hey friends 👋, after taking a few-year hiatus from social media, I've found myself lurking on Twitter recently. There is so much knowledge being thrown around, I'm excited to join the discussion. To everyone reading this, I hope all is well and the past year hasn't caused

[Jake Deichmann @jakedeich]

@ryanflorence What are you using to benchmark?

[Jake Deichmann @jakedeich]

@jasonlk It definitely goes wild on occasion, so you have to commit often, but you can also limit the tools it uses so it can't just run a database command. Worst case a file is corrupted and you just drop the changes.

[Jake Deichmann @jakedeich]

@reidhoffman We just launched @hellotwyn and now the AI advisors are going to start talking to each other.

[Jake Deichmann @jakedeich]

@ianDAOs We're building https://t.co/EPswS1RV6w, and decentralized identity will need to be a huuuge part for the AI agents to prove who they represent.

[Jake Deichmann @jakedeich]

@jazzplane @tbpn @realGeorgeHotz I started using Claude Code, canceled Cursor subscription. 0 switching cost.

[Jake Deichmann @jakedeich]

@gregisenberg We've got Claude automatically resolving 80% of user-reported bugs.

Bug comes in via intercom or email -> Claude to see if it's a question vs a bug -> Github issue -> Claude Creating a PR -> Posting the auto-deployed Vercel link and PR link in Discord to review changes. Insane.

[Jake Deichmann @jakedeich]

@michaelwatson2 https://t.co/C2HtGyOX66 does this. Let me know if you have any questions.

[Jake Deichmann @jakedeich]

@zachweinberg And this is a serious question. You're smart, and while you've hinted at your thoughts on policy, you've never fully expressed a vision for the US.

[Jake Deichmann @jakedeich]

@zachweinberg What do you think would be the best policy?

[Jake Deichmann @jakedeich]

@zachweinberg Or is it like the broken window theory? Fix all the obvious and crazy little expenses to stop them from piling up while they take the next 4 years to address the larger items.

[Jake Deichmann @jakedeich]

@realEstateTrent We built and AI that finds property owner contact data better than anything else on the market.

[Jake Deichmann @jakedeich]

@thesamparr Gotta feel bad to feel better

[Jake Deichmann @jakedeich]

@amasad @jeremyphoward What about programming robotic arms?

[Jake Deichmann @jakedeich]

@elonmusk what do you think of the show For all Mankind. Reminds me of SpaceX

[Jake Deichmann @jakedeich]

@erikzavodsky Do you even need a CRM? https://t.co/K3Qu0G0TJF

[Jake Deichmann @jakedeich]

@realEstateTrent

[Jake Deichmann @jakedeich]

@michelleefang

[Jake Deichmann @jakedeich]

@wagieeacc West coast is still coding and you’re asleep

[Jake Deichmann @jakedeich]

@JiffyLu @wagieeacc I had the same answer, but reading the question more it’d much more likely the bigger sample size will have n+1 boys to girls. The question isn’t asking about average distributions