This is really cool! I am curious how much is the average daily cost for the kinds of things you are doing. Are you using hosted models or running one locally?
Hosted models only. (my philosophy is that I need these things to be fast if I'm talking to them, and they also need to be maximally not wrong, which means cloud hosted big models even if they're expensive. To me, if it's wrong even once or if I'm sitting there waiting for it to reply, that's already making the value prop not worth it)
I think per token costs I calculated on Opus 4.5/4.6 were like $0.30/day for my text automations; $0.60/day for a few things I do that load up the browser. In general, anything browser-based munches up a lot more token (expected). What can be a bit of sticker shock is if you're having it load a lot of large web pages in a long conversation-- that can be several dollars. In the grand scheme of things, several dollars is not a lot but certainly from a "should I just go to the website myself" it tips the scale. I'm usually more interested in doing things once to "teach" it what to do (e.g. how to check a price) and then having it run that as a dialed-in cron job
i am actually very anti-notification (almost all are disabled; i use the iphone "notification summary" feature to get them delivered a few times a day only).
i do want to be on top of things and hate the feeling when i've double-booked myself or promised someone something that i forget. (i don't catalog everything in my fridge; the example was for a "deep freezer" i keep in the back room. this is super common for most suburban households but maybe lost on international audiences)
i am a normal level of busy and i think i was trying to have a bit of fun with creating things with a very low threshold for how good they had to be. i'm not bothered by everyone's speculations about me — i wanted to provoke some conversation — but this really is just about a bit of joy and building things i use every day makes me happy. people think it's cool that my calendar events make themselves.
it uses a real chrome browser window (that i can see when i remote desktop into the mac mini) that's been very good so far.
re: scale, is it that this whole project is worthless bc nobody needs it, or is it that its so good that scale is a requirement? this is a project i built for myself, i'm not commercializing anything
I use PurpleAir data for a lot of my home automations— I have a smart window vent and configure it to blow in/out depending on which side has the worse air.
(Thank you to those who maintain public sensors!)
I do notice that in my neighborhood (Noe Valley) a lot of the sensors are very incorrect or often offline. I've resorted to taking the median and throwing outliers away, but even this often doesn't work. This is the challenge of relying on crowdsourced data I suppose...
it's meant to throw out crazy low or crazy high numbers. i monitor 5 nearby stations and usually only 1-2 of them are bad (which is tolerable but not ideal).
Pay attention to the compatibility list though. If you use the Cloth™ on an iPhone 4 or earlier then you're in uncharted territory, who knows what could happen.
I thought the same. I had no idea there was a special cloth in the box. I'll have to dig it up and search for it. It sounds like a quality microfiber cloth is still fine though.
We agree! Unchecked development could lead to disaster. Insurers can insist on adherence to best practices to incentivize safe practices. They can also clarify liability and cover most (but not all) of the risk, leaving the developer on the hook for a portion of it.
1. Someone is always carrying the risk; the question is, who it should be? We suggest private markets should price and carry the first $10B+ before the government backstop. That incentivizes them to price and manage it.
2. Insurance has plenty of ways to manage moral hazard (e.g. copays). Pricing any new risk is hard, but at least with AI you can run simulations, red-team, review existing data, etc.
3. We agree on existential losses, but catastrophic events can be priced and covered. Insurers enforcing compliance with audits/standards would help them reduce catastrophes, in turn reducing the risk of many existential risks.
What you are saying makes sense for conventional harms like non consensual deepfakes, hallucinations, Waymo running pedestrians over, etc.
However, those are a far cry from the much more severe damages that superintelligence
could enable. All of the above are damages which already could exist with current technology. Are you saying we have superintelligence now?
If not, your idea of selling superintelligence insurance hinges on the ability of anyone to price this kind of risk: an infinitely large number multiplied by another infinitely small number.
(I realize my explanation was wrong above, and should be the product of two numbers.)
I think many readers will also take issue with your contention that the private market is able to price these kinds of existential risks. Theoretically, accurate pricing would enable bioweapons research. However, the potential fallout from a disaster is so catastrophic that the government simply bans the activity outright.
Dreambound creates upward mobility for American job seekers and simplifies hiring for employers. We’re an early startup helping low-income folks upskill and afford vocational training (think truck driving or nursing) with zero-interest loans. It’s a bit like Affirm meets edtech meets workforce development — deeply mission-oriented work helping people get ready for an economy that’s left them behind, and also an exciting engineering challenge as we’re working with a modern stack and awesome team.
This is engineering role #4 at a 10+ person company with some phenomenal talent and great culture. We're hiring for both front-end and back-end roles, with a focus on someone who has had 2+ years of experience working on front-end (React, TypeScript) or back-end (TypeScript). We have an SF office and a distributed team — on-site post-COVID and remote available.
In my time in tech I have never come across a company that is as mission-oriented as ours — while still very much focused on building a successful business. For many of our customers, Dreambound represents a lifeline to economic stability. We are backed by USV; leaders at Affirm, Twitter, General Assembly; and some of the best VCs/angels in the business. It's a GREAT opportunity to join an early-stage company.
You can apply here and mention HN, or email me directly (brandon@ the below domain name):
I still do all of the text automations, which have been pretty set-and-forget.
reply