ECC (and other methods) write the corrected value back to memory

That was my understanding (it corrects the error and writes the good value back to RAM), but now I’m not so sure! I imagine it must do that, otherwise a second bit flip would actually corrupt the RAM, and the RAM manufacturer would want to reduce that risk.

Regular ECC adds an extra parity bit for each byte. For each byte of memory, it can correct an error in one bit, and detect but not correct an error in two bits, so they wouldn’t want a one bit error to linger for longer than it needs to.

A better use of your time is to improve documentation. Developers generally hate documentation so it’s often in need of improvement. Rewrite confusing sentences. Add tutorials that are missing. Things like that. You don’t necessarily have to be a good developer or even understand the code of the project; you just have to have some knowledge of the project as an end user.

At least for that we have replacement names that make sense (like primary and secondary or replica).

HTML isn’t compiled, and unknown attributes are allowed. The best practice is to prefix non-standard attributes with data- (e.g. <div data-foo="test">) but nothing enforces that. Custom attributes can be retrieved in JavaScript or targeted in CSS rules.

I like the SwitchBot Lock because it sits over the existing thumbturn rather than completely replacing the lock. It still looks like a normal lock from outside, unless you get a keypad of course. I got one with a keypad so my dog sitter can come check on my dog and take her out while I’m at work.

It’s not internet enabled by default. You can buy a wifi gateway from SwitchBot, but instead I have mine connected to Home Assistant using Bluetooth via a Bluetooth proxy.

The whole point of Lemmy and the Fediverse as a whole is to decentralize, so everyone routing through a single service (Cloudflare) seems to go against that.

For DNS challenges, I personally prefer using acme-dns. It’s a separate DNS server that only serves ACME DNS challenges. I felt a bit uneasy using an access token for my actual DNS host since it grants full read/write access to every record. acme-dns reduces the attack surface.

Let’s Encrypt follows CNAMEs and supports IPv6-only DNS servers, so you could just run acme-dns on a spare IPv6 address (assuming your internet provider has a static IPv6 range, or you have a VPS with IPv6).

I think it’s all read-only access through, so you can only use it to import data not make new transactions.

That’s alright. Even read-only access is useful. I could write a script that pulls my current investments, prompts for the amount I’ll be investing in total, and prints out the buys (eg “buy 10 x VOO, 5 x VXF, 20 x VXUS”) that’ll keep the account balanced based on some percentages.

Does SimpleFIN use OAuth to log into bank accounts, or do you need to enter your bank’s username and password?

Unrelated to this post, but do you know if SimpleFIN supports investment accounts? If it does, it seems like an easy way to let me write a script to help rebalance my investment accounts. I might look into it.

I didn’t see a docker image

https://actualbudget.org/docs/install/docker

https://github.com/actualbudget/actual-server/blob/master/docker-compose.yml

It probably accepts other key types and it’s just the UI that’s outdated. I doubt they’re using an SSH implementation other than Dropbear or OpenSSH, and both support ed25519.

id_rsa shouldn’t exist any more.

Did you see something that said it was an LLM?

Edit: Indeed it’s an LLM. They published the model here: https://huggingface.co/reglab-rrc/mistral-rrc

Good idea to send donations to the syncthing-fork devs to keep it alive though.

In that case, could the syncthing-fork app be renamed to syncthing, now that it’ll probably be the main Android app for Syncthing?

mostly a wrapper around their proprietary library

I’m not familiar with exactly what Bitwarden are doing, but Nvidia are doing something similar to what you described with their Linux GPU drivers. They launched new open-source drivers (not nouveau) for Turing (GTX 16 and RTX 20 series) and newer GPUs. What they’re actually doing is moving more and more functionality out of the drivers into the closed-source firmware, reducing the amount of code they need to open source. Maybe that’s okay? I’m not sure how I feel about it.

Open source software doesn’t have a reason to lock you in like proprietary software does :)

More and more proprietary SaaS systems are allowing data exports now, to comply with laws like the GDPR “right to know”. Say what you want about Google and Facebook, but they were the first big companies to start allowing data to be exported before there was any law requiring it - Facebook in 2010 and Google in 2011.

Would you rather pay for every site you use? Not every site can afford to have someone else cover the cost for you (which is how Lemmy servers are ran for example), and the only other business models that have worked online are either running ads, or getting users to pay for access.

The advertiser doesn’t get your data directly. The data is aggregated using differential privacy so it’s impossible for them to see any individual’s data.

Manifest v3 is like activex. As of right now, it shuts down extensions they don’t want.

ActiveX was the opposite of this… It gave third party code way too much access. It was essentially unsandboxed native code running directly in the browser, like if you were to write a Windows app and let a site automatically download and run it.