Whenever this comes up, all I ever see are hand-wavy responses like this. "It does everything", "it changes your life." But how exactly? Because I can only see this being true for the terminally online.
That's fair, I'll give a couple of example of things that I've used it for...
I built an app (in codex), that goes through my emails, pulls out any receipts, and adds the purchases to a sqlite database. I do all my grocery shopping online (through my supermarket), so our weekly shop is stored in the database.
Hermes uses that to predict what items will need restocking (based on how often I buy it, and when I last bought it).
Every week, it logs into the supermarket website on my behalf, books a delivery slot, and adds all the items that need restocking to the shop.
Throughout the week, if I run out of something, I take a photo of the barcode and send it to hermes, it adds it to the shop for me.
I built a similar app for travel; if I have any upcoming travel, hermes will send me updates (delays, platform changes etc...). I do a lot of train travel for work, if my train is delayed, it will automatically apply for the compensation on my behalf.
It reads my emails, and pulls any actionable items out, and adds it to an Asana list. I can either do the tasks myself, or assign it to Hermes and it will do it for me. As an example, I had an task to book swimming lessons for my son, hermes monitors the swimming lesson availability and automatically booked it when a suitable one becomes available.
Shopping is a good use case, I have a clock that needed a new battery, it wasn't a standard battery, and required a screwdriver to open the compartment. I took a photo of the clock, sent it to hermes, and it found out what battery took and ordered it of Amazon for me.
I wanted to buy some smart thermostats for our radiators, but wanted to be compatiable with our current smart heating system. Hermes is connected to my home assistant, so it figured out what heating system we have and found a compatiable set of eBay (this is slightly more involved than it sounds, as there is a new version of the heating system with a new set of completely incompatiable accesories).
I don't think "it does everything" or that it has "changed my life", but all of those examples are chores, things that I don't particularly enjoy doing; and if something can take it off my plate, then it's useful for me.
It doesn't burn that many tokens, as most of the code is deterministic.
Receipt emails are in a predictable format, so codex built scripts to parse the emails. If there is a receipt email it doesn't recognise, it will use an LLM to build a parser for that kind of email once.
The agent also "self improves", so after asking once to get items to restock, it wrote a skill with the query it used in, so it doesn't burn tokens figuring it out again.
I was impressed that after it logged into my supermarket account, it inspected the network traffic and discovered a GraphQL API, and used that (instead of screen scraping); much more economic.
I've averaged 2.6 cents a day in token costs over the last month.
Thanks for your detailed answers. Although none of what you posted even remotely interests me as something I’d be interested in doing on a computer, it is absolutely fascinating to see details about how others are using LLMs and turning their computers into these “assistants.”
Yeah it’s not for everyone, it reminds me of the automation xkcd comic [1]
For me my motivation is that I’m working 9-5, my children go to bed at 8pm, and I don’t have my phone on me so I can spend time with them.
Once they’re asleep, I’ve got 2 hours to spend with my partner before bed. I’d rather actually spend time with her rather than checking if we have enough kitchen roll.
I understand what you’re getting at, but that example doesn’t bode well for you. The remote control increased convenience at the cost of making us more sedentary and lowering our attention spans. It was not a good trade.
I was thinking that I actually use grocery shopping as a recovery activity. I could cut it out but don't feel the need that much. It's not that I feel I'd do something incredible instead (which requires energy and/or recovery time of its own)...
Depends on how you define "reliably". You can get pretty damn close by triangulating on traffic patterns and browser fingerprinting. There is a lot of research in this area. But it'll never be perfect.
Interesting and thank you! Just like I'd like a "dumb" tv that I could just plug a Linux machine running Kodi into, I'd like a car that was 100% free from smart features, yet gave me the hooks to power any of the systems with my own software/hardware.
The closest thing is just any old basic obdii and that's it ICE car, where you rip out the oem radio and now have your power taps to do all sorts of fun hacker projects within that new found standard sized void in the dash.
Preorder customer and heavy info-digger here on Slate:
There's no actual wireless on the vehicle out of the box. There's a "secure gateway" module that has a USB Ethernet adapter hanging off it, then that hangs off a USB-C port just inside the dash panels. Job listings basically detail that it's going to be DTLS over USB to the truck to do all the CANbus magic and stuff.
The telematics module also just hooks to a second USB port on the secure gateway that also has access to that same DTLS link.
So far in the truck's development they don't even really have an app yet before launch but the truck's driving along very happy without it.
They say their vehicles have no modem by default, only leaving it as an option for fleet users. They also have no infotainment system to contain any sort of camera or microphone.
If you buy a completely option-free $24,950 truck, there is no telematics/surveillance in their vehicle.
Which also means there's no remote app features that people enjoy.
The spirit of the vehicle is you can option out just about anything and DIY-it; very, very few "from the factory" options, but dozens upon dozens of DIY accessories from the factory.
The current description for the module on their website:
> A plug-and-play solution that acts as a cellular bridge between your smartphone and your Slate vehicle, enabling features like remote access and battery pre-conditioning.
It's not clear to anyone exactly what features the module has to enable the above, or if it does more than all that.
Once the thing actually ships, there'll be manuals, and there'll be people hacking it to find out exactly what it both does and doesn't do.
And since it's a DIY addition, if you later become uncomfortable with what you added to your vehicle, you can remove it!
I do enjoy watching my friend remote start his Colorado through the app, it only takes 30 seconds for the app to load and sign in, and then he presses a button that used to be on the key itself.
IIRC, they don't have a cellular radio in the vehicle (or even a FM radio). All the connectivity is done via Bluetooth with a mobile app.
It's been a while since I read up on them and things might have changed, but I have a lot respect for what they're doing. Full embracing right to repair and customer privacy when the entire industry is moving at lightning speed in the opposite direction.
Aren't their companies that are doing their telemetry over bluetooth bridge to phone' internet? Not saying slate necessarily does, but it's certainly an available exfil surface, especially if there's a 'manage your slate vehicle' app.
Well, sure, but if memory serves, there are car platforms where you connect for say playing music and the cars app on your phone uses the bluetooth connection to exfil data to the car company via the phone's internet.
I'm just saying that "disable car's telemetry" is not necessarily 100% for preventing exfil of car telemetry if you then connect your phone while having the car app running.
Like, I'm not sure of the company, but I know there was some "I modded my car so it couldn't phone home but I'm still getting reports from the manufacturer about excessive acceleration" and it ended up being because the owner had disabled the car's telemetry but was using the app on the phone for remote start and music or whatever.
Can you explain your workflow a bit please? Do any of these tools work with Claude/Codex subscriptions or are they API only?
I actually built my own tool that maintains a work graph (DAG-like) with task leases. It allows me to copy and paste pre-written prompts into Claude Code, Codex, or OpenCode and all the agents self-coordinate through MCP calls.
I built this after trying hermes and Openclaw but not liking the lack of human-in-the-loop judgement. So I'm wondering if I should keep refining my tool, or evaluate something like pi?
Also funny that a plywood back is often preferred because it holds the squareness of the furniture without the issues of wood movement you get with solid wood backings.
Most days I feel like I must be the dumbest person on HN. I don't understand what 80%+ of submissions are about. But if it sounds interesting, I'll dig into it a bit and learn a few things along the way.
I don't ming a "ASK HN: What's your AI Setup", and then sifting through the 100s of comments on what people are doing. Then I'd scrape all that into my LLM WIKI and start digesting.....
I wholly agree with your comment, but is it legally "your code"? Copyright is implicit at the moment of human creation. But there isn't yet settled law on AI-assisted creation.
So it might be a problem for projects to accept contributions where it's not clear who actually owns that work.
Not a lawyer, but to some degree I think policies like the one Debian came up with can help the argument. You’re not submitting a stream of commits that all have an Authored-By: Claude Code footer and massive yappy commits and comments. The policy essentially forces there to be a degree of human authorship to each commit even if some of the bits came from LLM assistance.
Literally irrelevant and I have no idea why people even bring this up constantly. Copyright is practically unenforceable in the software world, when was the last time you saw someone (a company or an individual) actually sue _and win_? Apple wasn't capable of stopping all the other phone manufacturers from directly cloning their product (despite having patents & copyright) so how can an individual? The only thing that matters is execution.
For instance, my repos are cloned like ~100/week by what I presume are bots and AI lab scrapers using them for training material. How am I supposed to even find out who's doing it, and if they're abiding by the licenses I provided? It's impossible.
Lastly, having copyright isn't the same thing as being allowed to commercialize the software (and I believe loads of people confuse the two). You can still sell public domain code, even if you don't have copyright over it.
That might be your intuition but it’s not how the law works for code ownership. It’s very relevant for free software given the whole principle of free software is a copyright hack that protects a set of freedoms by relying on the copyright being recognized and enforceable.
If you drop the copyright claim you drop the protection of free software freedoms.
Software is one of the most derivative types of works ever invented. It's arguably based on copying what other people have already created. The modern world would look a lot more primitive if this wasn't the case.
Corporations like Coca-Cola long ago figured out that begging people not to steal your ideas is a waste of time. If you don't want people using your ideas, then don't share them.
Dunno why so many programmers don't get this and seem perpetually butthurt that other programmers figured out the ultimate way for programmers to rip off each other's code.
Coca-Cola literally shares their idea in every bottle they sell. The recipe being locked up in some super secure vault is a marketing gimmick and and has not relevance on anyone else being able to create clones of their drinks.
At this point, who could bring a copyright claim? Well, considering the massive AI training effort hoovering everything up, seems it might be approximately everyone.
Who would they be making the claim against? Well, considering the massive uptake of AI across the tech industry & beyond... approximately everyone.
Yes, the model companies are in the middle, but it's getting to the point where it seems a bit doubtful that claims will have any significant outcomes.
NYT might be able to get OpenAI to pay them for their content, but that's more of a financial tweak than a up-ending of industry.
Feels like those kinds of terms are moot in the big picture. Could they enforce it? Perhaps they can if a significant amount of the data produced is their IP, but who would they enforce it against? Consumers? Not much to be gained by going after them.
I can only think that they are hoping businesses might use the free plan and that they can sell them paid plans as part of an enforcement effort, but it doesn't seem like a very sustainable approach given the industry standard is to not claim ownership.
Not really. The term is a shortened version of "sucks eggs", which is originally from Shakespeare and used to describe an odious creature or habit (a weasel in Shakespeare's case).
