I saw the security article, but that sounds like it needs to be tackled by MSFT, the way Google has to handle Chrome extensions.
Have been a paid Jetbrains user for years, especially PyCharm. But recently, I had to do some front-end web development with ionic/Capacitor and Vue, and ionic only had a VsCode plugin. A few weeks later, came across Cursor which is a fork of VsCode with LLM support, and all the same plugins worked.
Still keeping my PyCharm subscription, but am wobbly on whether I'll re-up next year.
You all realize pretty soon no human is going to update an existing code-base?
Who wants to spend their time understanding 10 year old legacy code? They'll just feed it into an AI and tell it to add or fix a feature, then generate tests, and file a PR.
If it ends up having an airplane do a loop on take-off or sending your paycheck to Antarctica 🤷🏻♂️
Pretty funny seeing what is after all, a 'romantic' venture (dating apps) broken down in a completely utilitarian analysis. Also helps to have a brook-no-shit writing style.
Years ago, was asked to do dev work on a clone of a dating app. Ended up looking at all the things these kinds of apps did to get people to pay. Two features I remember were giving new users option to buy top profile placement for a limited time, so they showed up as a first choice for everyone for the next hour.
Another was sending paid 'virtual' gifts, which turned out were badly ripped off clipart of flowers and jewelry. This was when Tinder was first starting so none of the fancy retention methods described here or AI filters.
I passed on working on the app, but made sure my wife knew exactly why I had installed all those dating apps on my phone. It really was for research.
Tried it in Cursor (paying for both). 3/4 of the time get an error that says 3.7 is overloaded. Toggling back to 3.5 works. Will wait till the novelty wears off.
Just to be clear... I'm a massive Fediverse fan, and have concerns about BSKY's governance. But many communities streaming off Twitter seem to be heading toward BSKY because it's a shallower on-ramp.
Mastodon people recognize this and are working to smooth down the friction points.
What happens when their server expenses aren't covered, or bad people move in and every message has to be moderated, or the site moderators ban you?
And getting a whole community moved over... oof.
I moved a private mailing list to a WhatsApp group, then they changed their privacy policies. It took two years to convince people on to Signal, and 2/3 of the people didn't make the jump. And this was with a small group of people who knew each other IRL. Imagi e doing that for tens or hundreds of thousands worldwide.
This is why people are hesitant to get off Meta/Twitter. They're not going to do it again.
Your email server doesn't also run the group email list and all the join/drop/approve/ban operations. And if you bring your own email domain name, you can go somewhere else and get no disruption. But if you sign up for me@hotmail.com and hotmail bans you, you'll lose all your connections and conversation history.
The canonical list of operations on a social media platform far exceed that of an email service, a bulletin board, or a messaging service group. It's apples and rocket ships.
Bluesky is offering simple one-stop answers to a lot of these concerns. Fediverse needs to answer all these, plus address the whole long-term financial sustainability question.
The Fediverse experience starts with an unanswerable question: what server do you want to be on?
Most people will not have any way to answer that without knowing what the downstream impact will be. Mastodon people are working on smoothing that down, but it's still a pretty fraught question. And if half a given community ends up on one server and half on another, they get fragmented and conversations and followers fizzle out.
Bluesky wants to tell people they're not a single-node lock-in to avoid the Twitter effect, but it turns out that's their key advantage.
The only thing that will guarantee they don't end up like Twitter is if they revamp their corporate governance mechanisms, but they had to take VC money and haven't come up with a long-term revenue model, so it's not clear how they can avoid it.
Working on a new online gaming platform. There's internal debate as to whether to allow players to message each other and post in public spaces. It would make it a lot more fun for playere, but the risk of losing Section 230 'safe harbor' protections is a big concern. Also, the cost of moderation.
72 hours to finish a design project for a client. Hard deadline to make it to the printers in time for a trade show. By the end, I couldn't see straight. Slept for 18 hours after handing it in.
Saw the end product at the trade show. Horrified to pick out so many small errors. Customer was happy, though. Got two more projects.
Never did it again, but it was good to know I could.
A works/construction department in a medium-sized town. They had an Excel spreadsheet that had a HUGE number of screens. Anyone wanting to do commercial real-estate construction had to not only fill out these forms, but keep them uptodate and submit the updates at end of each work day.
The thing was HUGE and had lots of interdependent screens, where if you picked an item from a dropdown menu, it unlocked a bunch of other complicated screens or panels, and so forth. Each screen had 30-40 items and fields on it, and there were multiple dozens of screens you had to tab through.
To run it on the jobsite, construction contractors HAD to buy a pen and touchscreen Windows 'tablet' ($$$). The whole thing had been written and maintained by one guy over the course of a few years.
EVERYONE hated it. The guy who had written it wanted to get promoted to management, but nobody else wanted to maintain it so he was stuck.
Prediction: first big company to offer 32-hour work week at no loss of pay, with choice of remote or hybrid, will hoover up all the Grade-A talent.