If I understand correctly gzip, brotli and similar are best used to compress text.

Compression algos should be used on uncompressed data. Using them on already compressed data (most video, images, music formats) is generally useless.

Why 28Kb though?

It’s from an era where everyone just used open standards instead of trying to create their own thing for money.

SMS is literally from a time when every mobile phone manufacturer had their own charger plug. And some tried pushing proprietary headphone jacks.

Vendors LOVE vendor lock-in.

I am sober and serious.

At least it’s better than Slack and Discord. Still shit of course, but it could be worse.

Twitter currently has $1.5 billion/year deficit which is a lot, even for Musk, to bankroll.

Check Amazon’s Antitrust Paradox by Lina Khan (FTC). A very detailed review of how Amazon is a monopoly and how they dodge antitrust legislation.

That’s not how it works. Or, rather, that’s not only how it works. Sure, advertisers dream of users who see an ad once and run to buy a product. But ad effects are spread over time. They build brand recognition. They fake familiarity. Say you are in a supermarket and you want to buy a new type of product that you haven’t bought before. Very likely you’ll pick something familiar-sounding, which you heard in an ad. Ads pollute the mind even if the most obvious effects are, well, obvious and easily discarded, more subtle influence remains.

print("x") is you want to screw your students.

Because front camera is not just a feature I’m forced to pay for, despite never using it once in my life, but now it also makes my experience shittier by adding a hole in my screen. I hate that.

deleted by creator

We actually have a live experience of how that could go down

Another example: latest iteration of Google Captcha. Released with promises to end manually inputting text captchas, the main thing it turned out to check for is whatever you are logged into your google account. If so, you get through automatically, or, at worst have to press a checkbox. If you are not logged in, enjoy selecting fire hydrants and crosswalks.

Yea, it’s strange - top to bottom rows sum up to 94, 100, 92, 94, 100, 100

Research linked in the tweet (direct quotes, page 6) claims that for "GPT-4, the percentage of generations that are directly executable dropped from 52.0% in March to 10.0% in June. " because “they added extra triple quotes before and after the code snippet, rendering the code not executable.” so I wouldn’t listen to this particular paper too much. But yeah OpenAI tinkers with their models, probably trying to run it for cheaper and that results in these changes. They do have versioning but old versions are deprecated and removed often so what could you do?

cs.github.com is the best code search ever offered, especially at GitHub scale.

Email is not only 1:(small N). Maillists do exist and and are used to facilitate discussions between a large amount of people via email. They are also often public so anonymous readers and search indexers can use them.

/all is certainly an interesting thing - default Active sorting calculates a rank based on the score and time of the latest comment, with decay over time. If Threads are connected they would dominate /all. But there can certainly be adjustments, we can create a new sorting style, and make it default. For example:

• Posts are deprioritized based on MAU or some similar metric. The larger the MAU, the lower the post is ranked assuming the same engagement. If the post got 100 upvotes on an instance with 1000 users, it’s probably a much more interesting post, than the post that got 100 upvotes on an instance with 100 000 000 users.
• Posts are (de)prioritized based on instance source. For example setting Threads to -1000 would effectively remove it, setting Threads to -50 would allow you to see only super active posts. On the other hand if we want to see more content from less populated instance we might set it (i.e. german lemmy feddit.de) to the score of 100.
• Instances can provide a limited number or percentage of /all - i.e. after we got 10 posts from Threads, stop getting posts from this instance.

I don’t know. I would like to subscribe to someone on Threads from Mastodon (since both are Twitter alternatives), if they don’t have Mastodon account (which let’s be honest they probably don’t). Zuck does not get any of my data (besides what’s available publicly anyway). If Threads decides to go full EEE, I’ll stop getting updates from people on Threads, same as I don’t get updates from people on IG right now. I think proliferation of ActivityPub protocol would be the greatest advantage.

Moreover, I think we should follow the email architecture - I might use i.e. Proton Mail, but it does not prevent me from sending emails to Gmail, which I think is a bad provider, who collects a lot of user data. In fact if Proton Mail forbade sending email to Gmail I would be really displeased about that.

The goal is to allow people to choose where they want to go and ActivityPub is what can help with that, unlike blocking Threads.