One must always have main quests and side quests.
Main quest is doing your job, or taking care of your family, or both. Very simple.
But side quests? There’s so many options!
One of mine this year was to get over 100 interviews done.
And with a couple of weeks before deciding to leave Helsing, I had hit 102.
Oh yeah, I also left Helsing a few weeks ago. I want to take a month as a break to just chill and work on some personal projects while sleeping normal (by Mihail standards) hours again.
Given the amount of “Oh, just saw / heard that you left” messages I’ve been getting (about 2 per day for the last few weeks), there’s a high chance this is how you, former fellow Helsingeroo, find out.
Around late November I’ll look out for what the next great thing will be.
If you want to work with me, let me know what your quest is!
Stay tuned for eventual news on a new project of mine.
Most details available by following me on Mastodon ( kind.social/@Mihail ).
So, game I’m working on. Still being in “stealth mode” about the name and mechanics, but am happy to share that it’ll finally be a place where I can use my PhD research into game balance!
All the games I worked with for my thesis I had to fight to have parameters to change and metrics to compare against.
With my game? ALL THE DATA. ALL THE TIME. EVERY MORSEL OF INFORMATION WILL BE FED TO THE GENETIC ALGORITHM.
The screenshot is less than 10% of the scope of data available.
I really really really care about the quality of the products I use. And that transfers to the products I am working on delivering.
As a result, I care a lot about testing strategies. Especially performance testing.
So I wrote down my current views on the topic, for anyone curious, or anyone wanting to learn some basic concepts.
1 like 1 repost
@Mihail@kind.social shared this @Mihail@kind.social liked this
Already counted in the numbers above.
I hope everyone’s had a very relaxing week. Especially IT admins and security experts.
1 like 1 repost
@Mihail@kind.social shared this @Mihail@kind.social liked this
Already counted in the numbers above.
You can clearly tell there are multiple teams not communicating on what features they are delivering.
When’s #vutuv adding “N people viewed your profile. Wait, no, N*2+7 people. Wait, actually it was 12 bots in a trenchcoat and your ex.”?
Oi, LLM, implement this feature for me.
“OK!”
…
Oi, LLM, why did you use String for the dates?
“No strong reason. You are right to note that. A DateTime object would be a much better fit and be less error-prone.”
…
Aaaaaaaaaaaaaaaaaaargh.
This is why I read the code. And write tests. And benchmarks. I do not trust the stochastic parrot.
HA London TfL automation of the day:
This week’s been fun already, and it’s only Tuesday. Multiple issues and PRs raised by contributors for both gathers.cards and HA London TfL!
Great feedback, good errors caught, happier better software as a result.
For anyone considering gathers for their collection management, do discuss how you’d like a deck-specific view / functionality to work - codeberg.org/morosanmihail/…
A shame I won’t remember this given the cold that hit me overnight.
I am entertained by Shizuku for Android. It lets you do advanced operations on your phone, which usually require adb from a PC.
It does so by you enabling Wireless Debugging on your phone, then Shizuku, on your phone, remotely over the network connects to your phone, then starts a local version of adb on your phone, so it can remotely connect to your phone and do stuff.
…
Silly, but it’s what you need to do in the age of walled garden Android.
It’s good though. github.com/rikkaapps/shizuku
I love open source. It has made my projects a lot better.
Someone found a bug? One of three things will happen: they just accept it (the status quo), they raise an issue (lovely! Gives me information to help solve it), or they just put up a fix themselves (lovely! Gives me the fix).
With traditional closed source software you barely have 2 out of those 3 options. And that’s if you’re lucky and they have a proper way of raising issues directly to developers and not through some labyrinthine customer service. Which might, or might not, be just an LLM bot pretending to try and help you.
Yes, a few peeps found some bugs in the Home Assistant integration I maintain and fixed them. Wonderful people!
3 likes
@diwesser@mastodon.social liked this @Krisuuu@mastodon.social liked this @tobiaskuenning@mastodon.social liked this
Already counted in the numbers above.
9/10 entrepreneurs say that replying to this post with your favourite thing that happened in the last few weeks is a great idea and will have massive returns.
Who am I to disagree?
I’ll go first: got to see my amazing wife rescue a tiny little bundle of orange feline joy.
After a bit of a holiday, decided to put some thoughts I had on working in defence to text.
Released v0.7.0 and v0.7.1 of #gathers yesterday and today respectively!
Plugin support is landed and now working on a bigger demo plugin for book management.
Also further improved performance of the MtG side of retrieval. Lots of performance was left on the table due to missing indices.
Enjoy! I know I will. Because I’m my favourite client. codeberg.org/morosanmihail/…
Finally on holiday that’s not at a beach. Which means I can spend a bit more time on Gathers.cards. And finally attempt to solve “how do I handle various foilings for various cards across multiple card games?”.
And… just maybe… also expand the plugin system to have a book collection as well. Gotta make sure I don’t get duplicates of a book.
The European Commission is funding a European project to develop a safe, inclusive, and decentralised social media platform for young people. The platform must be based on ActivityPub or the AT Protocol and co-created and tested with young people from at least eight EU Member States.
The project will focus on privacy, mental health, user autonomy, moderation, the responsible use of AI, and a sustainable adoption strategy. The Commission expects to fund a single project with up to €1.48 million, covering a maximum of 80% of eligible costs.
As an open and decentralised alternative to LinkedIn, operated in Germany, vutuv could be a good fit for this call. I am therefore considering whether I can find suitable partners to form a consortium.
The consortium must consist of at least four independent organisations from at least three different EU Member States. They do not all have to be companies: NGOs, public bodies, and other legal entities are also eligible.
Can the Fediverse—that means all of you reading this—help me find suitable partners?
I am particularly looking for organisations with experience in youth engagement, software development, mental health, digital inclusion, media, community building, or scaling commercial or non-profit services.
The application deadline is 6 October 2026.
Would you or your organisation be interested in participating? Please contact me as soon as possible. Recommendations of suitable organisations are also very welcome.
Call details:
ec.europa.eu/info/…
8 likes 13 reposts
@Friesenloewe@mastodon.social liked this @SuperMoosie@mastodon.au shared this @MrLee@aus.social shared this +18 more
Already counted in the numbers above.
Smuggled myself into BIGGICon last night. Well, smuggled is a strong word, given I was invited, but had not registered because I didn’t think I could go.
…what’s BIGGICon? Time for a brief history lesson: the PhD scheme I did was called IGGI and I was part of the first cohort, 12 years ago. IGGI (Intelligent Games and Games Intelligence) is now almost done. Over. Last cohort was onboarded 3 years ago (all planned, by design and funding). IGGICon was a yearly event showcasing the students, their work, and the industry partners. BIGGICon is just the last and biggest IGGICon.
While academia as a whole left a sour taste in my mouth, despite an amazing supervisor and a topic I really loved, the people in IGGI are amazing. And the people supporting the students from industry.
So next time you see someone boasting an IGGI doctorate or experience, they’re likely amazing and you should hire them / give them your undivided attention / donate all your wordly possessions to them.
Oh, and I’ve been told The Chinese Room has some amazing announcements soon (TM). Looking really forward to that.
New feature I’m manually testing for #gathers : plugins!
My solution for enabling different collection types and items, but not forcing people into recompiling the Rust code, or having to create new forks of the server, was to enable a REST API and the main server can point at as many of them as you want. Then you can host them however you want!
Which could enable some cool new use cases for Gathers… though I am not sure what a Foil copy of Dune is meant to be… a hardcover?
We can figure out good solutions for all the technical problems. And the trust part is more important for users than for server owners. For the server owners this would be something like an insurance.
Hmm, you do make a good point. The average fedi user would care a lot more about where their data ends up than the usual social media user.
I don’t think I’ve seen this “self-healing” approach to services yet, but it would be amazing.
It obviously brings a lot of concerns (domain ownership, data backup and exfiltration, how you handle updating your safety net servers if those close first, etc).
But the benefits are definitely there and definitely wonderful. But it’s clearly a server and infrastructure problem that might force some “if you want this feature, you must deploy this way” requirements.
(And yes, these are technical problems, but they would directly influence some server owners in whether they trust a feature like this at all)
Auf der Rückfahrt vom FediDay hatte ich noch eine Idee.
Ein Problem ist ja, dass (gefühlt) immer mal wieder kleine Fediverse-Instanzen schließen. Aus welchen Gründen auch immer. Das wirft ja ein denkbar schlechtes Bild auf alle im Verbund. Bei den Banken in Deutschland gibt es ein System, das die gegenseitig für einander einspringen, falls eine den Bach runtergeht. So was Ähnliches könnten wir im Fediverse doch eigentlich auch machen. Sprich, wenn einer kein Geld mehr für eine kleine Instanz hat, die Hardware den Geist aufgibt oder sonstwas, dann könnte er sich an diesen Verbund wenden und sagen: „Übernehmt bitte xyz.social, hier ist die Domain”.
Bitte jetzt nicht als erstes mal technisch lösbare Probleme wie Backups diskutieren! Die Frage ist, ob WIR - also die Server Betreiber - so was prinzipiell machen würden oder nicht. Ich hätte keine Probleme damit, ein paar kleine Instanzen auf den vutuv.de Servern ein Gnadenbrot zu geben. Ich will nur nicht der Einzige sein. Das muss schon fair unter allen aufgeteilt werden.
Was meint Ihr? Wie gesagt: Bitte erst mal das “ob überhaupt” diskutieren. Danach kann man dann über die technische und organisatorische Probleme sprechen. Wo ein Wille …
55 likes 7 replies 54 reposts
@BernhardHu@mastodon.social liked this @dkl@23.social liked this @vhs_berlin@vhs.social liked this +106 more
Already counted in the numbers above.
I don’t think I’ve seen this “self-healing” approach to services yet, but it would be amazing.
It obviously brings a lot of concerns (domain ownership, data backup and exfiltration, how you handle updating your safety net servers if those close first, etc).
But the benefits are definitely there and definitely wonderful. But it’s clearly a server and infrastructure problem that might force some “if you want this feature, you must deploy this way” requirements.
Given my strong opinions on data sovereignty, quality software and avoiding orphan-crushing megacorps, you will not be surprised to know I heavily dislike Meta, Microsoft, Tencent, Amazon, etc.
What you will be surprised is that there’s also open source projects I heavily dislike. And that’s usually because of the people behind them. Evil people are evil even if what they release is open source. Giving money to evil people with evil agendas will, drum roll, lead to evil.
So while I could personally type all the reasons I dislike Omarchy and the ghoul behind it, this link can do it for me a lot better with even more info. Do I agree with every single point in there? No. But I agree with a high majority of them. stopomarchy.neocities.org
Helsing’s own Jon Gjengset (also known as the author of Rust for Rustaceans) will be speaking at bcnrust in October! eurorust.eu/activities/…
If you know me, or read some of my older posts on other platforms, you may remember I am a big advocate for “stop using technology from companies that are making the world worse”.
That includes reducing reliance on Google, Amazon, BP, Shell, Meta, TikTok, and many more.
And definitely to stop using their software and hardware, whenever possible.
So if you were unsure if it’s worth the effort, take a look at this article ( desmog.com/2026/… ) and tell me if you want to keep giving money or data to these organisations.
Because the money you give to Google will reach Reform (if you’re UK), christofascists, and it will slow down clean energy initiatives.
And if you want to support those, please just block me.
3 likes 2 reposts
@thisislieven@c.im liked this @RichardoC@infosec.exchange shared this @drathir@mastodon.social liked this +2 more
Already counted in the numbers above.
