The image is a screenshot of a Twitter post. The post features an illustration of a character with turquoise hair and teal eyes, lying on her back amidst a cityscape. She is wearing a sleeveless top with teal accents and black pants. The background shows a dense urban environment with numerous buildings. The tweet is in Japanese and includes hashtags related to the character's birthday. Below the illustration, there is a reply from a user named Lucario Remendez on YouTube, expressing a comment about the character's age in the art. The reply mentions that the character has turned 18 but will always be technically 16-17 in canon, and criticizes the art for depicting her in a mature manner. The tweet has received 488 retweets and 3.7K likes.
―
The image shows a Twitter profile page for a user named Lucario Remendezo on YT, with the handle [@]FuecocolsBetter. The profile picture features an animated character resembling a Pokémon, specifically Lucario. The background image shows a character from a video game, possibly Fortnite, holding another character. The profile description reads, "My Bum Is Really Fuzzled 😂😂😂 #1 Hatsune Miku Hater [@]lolicollection_ got their ass banned yipeee chapter 3 is garbage." The profile includes a location, "69 Cum St. Your Mom's House," a birthdate, "Born September 3," and a join date, "Joined February 2025." The profile has 999 following and 114 followers, with a note stating, "Not followed by anyone you're following." Below the profile description, there are tabs for "Tweets," "Replies," and "Media," with the "Tweets" tab selected. A pinned tweet from July 13 reads, "Replying to [@]BackersGamesF: Miku fans are retarded," with 11 replies, 3 retweets, and 12 likes.

Provided by @altbot, generated privately and locally using Ovis2-8B

🌱 Energy used: 0.826 Wh

Das Bild zeigt eine Szene an einer Brücke, auf der ein Straßenbahnunfall stattgefunden hat. Im Vordergrund ist ein beschädigter Straßenbahnbahnhof zu sehen, mit mehreren Personen, die sich um den Unfall herum versammelt haben. Ein Straßenbahnbahnhof ist stark beschädigt, und die Straßenbahnen sind auf den Gleisen verkeilt. Auf der Brücke sind mehrere Autos zu sehen, die in beide Richtungen fahren. Einige Menschen stehen auf der Brücke und beobachten das Geschehen. Im Hintergrund ist ein Fluss zu sehen, und am Ufer sitzen einige Personen. Die Umgebung ist grün und es gibt Bäume und Gras. Die Szene wirkt chaotisch, mit vielen Menschen und Fahrzeugen.

Bereitgestellt von @altbot, privat und lokal generiert mit Ovis2-8B

🌱 Energieverbrauch: 0.190 Wh

Captured with a small mirrorless camera and 18mm lens, on a small tripod. I had no shifting lens, so this was shot head on with a wider lens and then cropped.

Harsh midday sun is normally not ideal for architectural photos, but the hard light nicely brought out the clapboard siding of this (since demolished) rather dilapidated old former antique shop.

@purplepadma

it sounds like they've at least got a plan? i hope you're spared any more Big Ow though anyway.

I realized later that basically I would always do the test cases manually one time as I wrote the code. Like each little part I wrote I would check by running the binary with special inputs and debug prints etc before moving on to the next part.

So writing a test is basically like writing down what you would have manually checked. Because it's code then you can do every check every time.

Later on in a project when you refactor it's easy to forget some things you should check in old systems and it can become "I have to carefully go through the code and check re-check every branch to know that this works".

Having tests really makes refactoring internally and not having your API break feasible.

It's important to write tests thoughtfully though and really target the different possible branches in the code. It worries me a bit when people say "I just tell the LLM to generate some tests and it works great!"

One thing I learned from work is "only test the public API" which is a way to avoid "change detector tests'. I also always try changing the code to break it and make sure my test actually catches it.

Tests are mainly useful when they don't change while you change the underlying code, to verify that the behavior hasn't changed. If you test the internal parts too much then you lose the ability to refactor and keep the tests the unchanged. At the same time it can help to have some key internal tests to diagnose which part of a big system is failing. There can also be internal structural APIs you want to test the "public" interface of.

The point of a test is to make sure something important doesn't change.

And what do you do about the odor evidence? Spray bottle of bleach?

Just admit you crank it in the study and leave off with the deception. Also, never attempt a secret affair, as you'd be caught out in minutes.