Thread is great too. Z-wave and thread are quite similar in terms of power consumption.

I literally just bought a satellite 1.1. I bought an atom echo to POC whether I could get reasonable wake word + stt + LLM running locally with good latency with the intention of building my own speaker afterwards.

I went with the satellite 1.1 because:

1. I haven’t done a good audio driver circuit since university
2. Getting 2-3 PCBs printed and SMD soldered sounds like a PITA
3. I didn’t want to design an acoustic cavity
4. Satellite 1.1 w/ xmos is just really hard to compete with, and the on-board humidity and temperature and stuff removes the need for me to have a distinct sensor for those in that room

It’s a bit more expensive but I’m really optimistic. It hasn’t arrived yet, I’ll lyk how I like it when it is delivered!

I saw a project (echomuse) that converts an EchoDot gen 2 into an HA speaker that I thought about doing as well but I didn’t really wanna spend $100+time when I could spend $200 and no time for a better product and support a cool company.

Jesus, that’s bad. I just installed an Ultraloq Bolt SE and it’s still reporting 100% charge after 2 weeks of daily use. I bought mine explicitly for battery life, I hope I don’t have the same thing happen…

Search should be a public utility at this point.

But then again I don’t know how you’d neutrally maintain a search algorithm.

Regardless, depending on a private company for something so fundamental to societal functions seems bad.

So… I’m with you/them about the vendor lock in but I think they’re slightly underestimating what ollama brings to the table. For example, ollama bundles CUDA libraries so I don’t need to worry about that as I wrangle with my GPU. It’s also well supported by other ecosystems like home assistant in a way that raw openai compatible API server isn’t.

Overall I don’t know if I find these arguments compelling at this time.

Man you can absolutely run something on ollama yourself. Right now I have ollama running qwen 3:4b-instruct on 6vCPU and no GPU and for very simple conversations it’s quick enough. TTFT is slow when prompt is uncached but otherwise TTFT is like 2s and I get maybe 5 words/s out.

If I gave it my 1660 I could probably get ~100tps but my card won’t pass through to my VM properly…

Most of this continent was built by racists who owned car companies. Turns out racists that own car companies really wanna make sure that everyone has to drive and black kids stay away from white kids, and then they build cities to make those things happen.

I mean, yeah. Asking “do xyz make no mistakes” won’t get you what you want. Come up with a plan, decide on what constraints matter, pick libraries you want to use. Decide what doesn’t matter. Generally around now I feed it my plan and ask it “do you think I’ve missed anything?” and read what it suggests and see if I agree with any of it.

From there, plan out what incremental progress looks like. Maybe start with the UART/I2C/SPI library and just reading firmware version off of each child component. Then ask it to go through the plan, increment by increment. Review everything as it does it, watch the console and redirect it if it’s making poor choices or just things you know will box yourself in in the future.

On my best day in university my project this weekend would have been weeks of work. It took me a day, not including soldering and packaging after the electronics were functional. It’s just undeniably faster at this point. Even if the only thing I used it for was to open the serial port, watch the logs, and fix any bugs I introduced, it would be faster. But you can feed it whole datasheets and it’ll construct a client library for your device essentially perfectly on the first try.

I have serious ethical qualms about how LLMs were created, how they’re going to be used, the inequality they create, the finances of the industry behind them, the environmental impact… but from a purely technical perspective they can be very powerful tools. But you need to already have domain knowledge to be able to massage it properly - “everyone is engineer” still isn’t reality. Someone with experience and good intuition though…

I dunno man claude one-shotted a firmware build for an ESP32-based project I was working on last weekend. I had to massage it properly and I wouldn’t have been able to supervise it properly without my EE degree but it did produce valid and correct firmware.

So in his case he ended up replacing the battery because he couldn’t source a replacement part he needed, but he’s still got the old battery in case he can get his hands on replacement contacts. Sounds like he wouldn’t recommend the grease approach, it did buy him some time but also caused a worse problem later via the arcing. He also said he did maintenance on the battery lock/latch to get it to jiggle less.

Sorry for not having more specifics!

My buddy had to replace the contacts on his battery and ebike this past spring after he absolutely ruined his contacts with snow and gunk. He was able to make due initially with dielectric grease and a contact cleaning but eventually he jiggled too hard on a bump and sparked the contacts. Let me ask if he repaired his battery or replaced it. He had a stock Aventon Level 2 battery from 202…1?

Whois still has the DNS record but it doesn’t resolve for me either at this point. I can’t resolve it using dns.leonet.it either so presumably the sanctions have taken effect and their internet service has been cut off.

May I ask why FB, Twitter, etc. get protection from liability from 3rd party content under Section 230 of the Communications Act but these folks don’t? They both have a TOS and reserve the right to refuse service.

I’m not downvoting you btw, I know you’re engaging in good faith. Idk who is.

So they have a TOS like Facebook or any other large platform, and reserve the right to refuse service?

Unless they’re knowingly doing business with a known convicted terrorist, I fail to see how this makes them a terrorist support platform.

Anyways in the US Section 230 of whatever law prevents companies and organizations from being liable for content users post to their websites. How is this any different than FB hosting the breeding grounds for Proud Boys et al, other than them not being a US company and them being leftist-aligned?

I’d argue it doesn’t matter. If they offer e2ee email, they can’t be aware of the content of any messages exchanged.

Unless they were informed one of their email accounts belonged to someone convicted of committing an actual act of domestic terror, they would be completely in the dark about who’s using their email services.

Given its this US state department issuing the statement, I wouldn’t believe any accusation leveled at all without proof. I’d need to see evidence that A/I knowingly provided services to a convicted domestic terrorist for me to believe this is anything other than trying to stamp out E2EE.

https://autistici.org/who/manifesto

Seems pretty great to me. MSFT and friends must be unhappy Europe is slowly moving off US software.

This is conjecture because I can’t access the full paper at this time, but based on:

The learning losses are concentrated among roughly 80% of AI users whose behavior is consistent with homework outsourcing, as indicated by exceptionally short homework completion time coupled with high homework scores.

I’m guessing this is an artifact of “students under higher pressure to perform are more likely to use AI more thereby lowering their scores”. Given Chinese patriarchical cultural biases, I’d imagine boys are under higher pressure on average, thereby leading to greater usage.

So perhaps it’s actually the second one, as the abatract is unclear if they’re studying performance relating to usage or not. Without checking methodology, it’s unclear if they’re controlling for usage.

I suspect it’s the second but now that you mention it, it is a bit ambiguous. I’m assuming the study would control for usage to compare outcomes across similar usage metric cohorts.