Notices by Mario Zechner (badlogic@mastodon.gamedev.place)

The web can be a cool place.

Jacob from Toronto has been inspired by my grocery price data mining bullshit and started scraping Canadian stores earlier this year.

He has lots of data, and needs help. Want to do something fun, educational and possibly useful?

Here:
https://jacobfilipp.com/hammer/

I've incorporated your feedback in my Mastodon thread reader @mastoreaderio

- It honors #nobot in your profile and will not unroll your threads. The requester gets notified via DM.
- It supports unrolling via DM. When requesting an unroll, set visibility to "Mentioned people only". Keep timelines and threads clean! It will remind you to do so.
- Otherwise, its replies will have visibility "unlisted" to keep the local instance timeline clean.

Special thanks to @untitaker for his input

It appears Google's AI overview is really just RAG over the top search results (= search stuff on the web given the user query, let LLM answer query given the contents of search result)

Naturally, redditors have started poisoning search results. Google ranks Reddit very highly. This issue is basically not fixable unless Google largely ignores Reddit content except for known good content, or develops a model they can use to filter adversial content (which is basically impossible)

Glorius.

And if you think perplexity.ai is any better, think again. They are all doing the same shit:

- Search top web results for query
- Stuff results plus query into a prompt
- Ask LLM to answer query based only on contents of search results.

Tada (Note: the answer looks superficially better until you get to the coconut section :)

We have an extreme right wing media outlet in Austria called Exxpress. They have a website and a "TV channel" that (used to) broadcasts on a local cable network, as well as on YouTube.

They recite Russian propaganda verbatim sprinkled with the same old extreme right wing narratives.

They also got €1,000,000 in state funding.

I decided to have some fun and did a thing.

You see, they embed the YT Livestream on their website. It autoplays, withiut consent. The views there are counted by YT.

Help needed. Can you all open this link, play around a little, then reply with your device/operating system/browser info an how kany "fps" you got?

https://marioslab.io/uploads/webgl-demos/example/physics2.html

You'd really help me out!

Well, Mastodon is even more fucked than I thought when I built my own client.

Turns out an instance may eventually evict posts from accounts from other instances.

Say you are on instance A, and a friend is on instance B. You two have a convo. After some time, all of B's replies are gone from your instance. If you view your thread, you are missing all their replies. If you are lucky, instance B still has all replies. Or not. Poof.

What a terrible, terrible platform.

https://mastodon.gamedev.place/@rovarma/111348337555786854

@z428 @HistoPol by design

Quite a few people sure were surprised to learn that all their posts can be read without a BlueSky account. Especially those who fled Xitter to be a bit more sheltered from harassers and nazis.

Welp.

For all its faults, Mastodon is the better platform for people who care about having control over their data. It's definitely not perfect in that regard either, but at least it tries.

Also, it has GIFs and videos and polls. Based on BlueSky's dev velocity, they'll have that sometime 2032.

The past two nights I wrote a "thread reader app" for BlueSky.

https://skyview.social

Oh boy. The protocol is absolutely insane. RPC galore, responses are only partially typed. The docs are pretty much useless.

But the "funniest" part is this: there's no privacy. And I don't mean missing DMs.

All your posts are available through API endpoints. Without any authentication. By design.

The "invite-only" thing may have you think otherwise.

Here are my last 100 posts.

https://bsky.social/xrpc/com.atproto.repo.listRecords?repo=badlogic.bsky.social&collection=app.bsky.feed.post

Letting the little one play on the office desk was my most brilliant idea as a father so far.

Follow me for more parenthood tips.

Introducing Masto Reader. Like Thread Reader App, but for Mastodon.

How does it work? Mention @mastoreaderio(@mastodon.social) anywhere in a thread and say "unroll" and the little bot will do the rest.

Here's an example (that takes a bit to load for reasons)

https://mastoreader.io/?url=https%3A%2F%2Fmastodon.gamedev.place%2F%40badlogic%2F111071396799790275

You can share those links with others.

Links to the crappy source code can be found in the bot's profile.

Everything runs in your browser. The server doesn't log anything either.

Enjoy and spread the word.

@kentindell iirc, the UK sources a lot of food from countries like Spain. With Brexit, that became significantly more expensive. Not an expert on that tho, take this with a table spoon of salt.

@kentindell where "investments" are things like building new stores or parking the money in financial assets. Austria has the highest density of grocery stores in all of Europe. There's absolutely no need for this density.

This little trick allows them to lower their official "profit" numbers in their financial reports.

The UK fucked itself with Brexit when it comes to food prices, so your situation may actually be caused by the price increases of out of country producers.

@kentindell my data can not on its own. However, for basic groceries, i.e. milk, specific meats, some fruits, another governmental org recently released the prices the grocery chains pay the producers. In that case, it's the stores that did the gouging.

We do observe stagnant/decreased year-over-year profits for grocery chains in Austria as well. But there's a big but :)

You see, their operating expanses have not increased dramatically. What has increased are their "investments".

@iworx no need for protection. The thing is easily self-hosted. Can even deploy it as a GitHub site, with a daily workflow run scraping the data.

@DavBot I'm all for that. That's why this is OSS and easy to self-host. Adding a new store is usually less than 200 LOC (code quality is terrible tho :D)

https://github.com/badlogic/heissepreise

(No the code isn't great, and I should have used TypeScript. I didn't plan for this to become a thing. I'm sorry)

@swart if it can be displayed in a browser, it can be scraped. Some are obviously more easy to scrape than others. Walmart CA seems to be one of the easier ones for example.