pianoplant @ pianoplant @lemmy.world

Posts

6

Comments

44

Joined

2 yr. ago

This seems like very standard ML. I'm not surprised it works, but also it likely takes a huge amount of training data (i.e. print samples) to recognize a specific machine.

I've done stuff like this. For instance I took a pre-trained model that could identify animals and used reinforcement learning to feed it thousands of annotated images of my cats. After this fine-tuning it could reliably tell the difference between them. Useful? Yes. Neat? Yes. But it's not like it can identify a cat it's never been trained on.

So it's interesting and useful, but not as impressive or useful as the article makes it seem.

Also I'm sure something as simple as changing a nozzle or even what slicer is used would completely throw it off.

I really like this game. Agreed with the other comment about the endgame being a little weak but it's still early in development and the content they've added has been great.

I don't think this is that unpopular. But it's probably better to be radically honest with yourself. Notice and accept your flaws, yes, but also notice and accept your value and strengths. And if some of your flaws are reasonable to put some work into then you may wish to choose to do that. For example - I'm trying to be more kind this month.

Ironically AI would probably work ok here whereas excel is using a human-designed pattern matching heuristic that apparently either has a bug or didn't take into account your locale properly. I say that as someone with a relatively negative opinion of ai

This is a task ai would do well at whereas most of the Excel workflow it would not.

It's actually because January is also misspelled. Or possibly because Excel's language/region isn't set up right (I see your month abbreviations aren't us-eng... If excel is in us-eng it likely isn't going to identify them properly.

All of his books are just wonderful.

If this was real they'd definitely say 0 fat, 0 sodium, 0 sugar. Companies are allowed to round down under a certain amount un the US

Shirokuma Cafe (polar bear's Cafe).

Paprika. To see the parade for the first time again... Wow.

Keio line is in Japan. So technically it's a train-nigiri.

Sadly about 50. But that's all my plan allows at my address so not a hardware issue.

I'm on 1.2gbps with my own modem... That's the fastest available at my address.

Not what you're asking but since it's been covered well:

Buy your own cable modem and put your own firewall behind it. Not only will this save you money in the long run, you'll also have no issues with things like port forwarding. I use Comcast/Xfinity with a docsis3.1 cable modem + a decent firewall and it's a good way to go.

Paperless-ngx installed via docker-compose is super easy. I have it on a luks-encrypted vm only accessible via tailscale.

I've been trying to move away from email as a document server.

Anything that's important / I might want to reference later gets exported to a secure paperless-ngx instance where it's neatly categorized and easily searched. I then delete it from my inbox.

You're thinking of profit. Revenue is all money coming in before expenses. Revenue is still a big number even if they're losing money.

I love the listicles feature. Not because I want to actually look at those results but because they're all stripped out of the main search. Doing product research is so much better without all those low-quality "top whatever 2023" lists that take up the first 3 pages of google lol

There are a lot of options: Bitcoin & Bitcoin lightning via opennode + Paypal, Venmo, iDeal, giropay, & stripe.

Personally I converted some XMR to BTC then paid with that (using a single-use wallet)

Kagi at the moment. Just feels right to pay for incredibly useful services with money instead of privacy

By default just did a video (piped link) on this and I 100% agree with him. The killer feature is simplelogin. Being able to use a different alias email for every single account I use is absolutely amazing.