I want to automate the first half of software engineering work.
I'm building a system that reads Slack, listens to Google Meetings, user complaints, etc and gives me prompts I could feed into coding agents or planners.
Problem-to-prompt seems like a larger obstacle than coding these days, I wonder if it's solvable, and if solving it makes cheaper coding agents viable.
It has to be the auto-playing Tomb Raider agent, where LLMs were used to give Lara self-awareness. I've never seen anything like it.
It starts off with some classical computer vision shenanigans to understand the character movement, map layout, and to create the 'desire' to explore. Then the LLM is given input of images, sound descriptions and prior thoughts, lettting Lara remark on the situation, which feels very surreal and, at least for me - very unexpdcted. E.g. she hears the wolves howl and wonders how they survived in this environment. Or meta-remarks on game music changes.
Worth pointing out that most of that video is fake[1]. Though it and its debunking video are still a great example how to make entertaining fictional content with a little help of AI. It probably won't be too long before somebody builds something like that for real, similar AI mods[2] for Skyrim are already out.
Ligatures became one of my favourite features after trying out Fira Code. It felt like an obvious improvement that I was yearning for, for years.
I guess my brain likes having distinct continuous symbols to represent different operations - it reminds of of math in school & university. And I don't see a problem during editing, knowing they're made of multiple symbols.
And yes - I feel that I do struggle much more without them.
Completely out-of-the-blue, unproven guess - my mind is used to learning new symbols quick, from all the gaming I've done. And it's much easier to learn a new symbol that's cohesive, continuous & unique, rather than having to read disjoint characters and figure out a different meaning for them.
In Edinburgh (2013) we did get Haskell as the very first language of the very first course (yes, functional programming before OOP!). Having coded before, I was pretty confused, why would they do that, but it ended up being an amazing equalizer. We all started at similar levels of understanding and ramped up together.
So, here's the thing - there's a secret world of sports and odd activities that, I feel, only a select few know about, that are both more fun, cheaper, and more beginner friendly than mainstream stuff.
I found, a while back, that things like Lightsaber Fencing, Megagames (board games with 60+ people), dodgeball, Historical Fencing, Ultimate Frisbee, Apenkooi (tag, dodgball, minigames in the Netherlands), etc - I found that those exist. And I finally feel great doing sports - it's great to explore around a bit, do something unusual for once.
The issue is - communities are small, barely known. And it's a wee bit tricky to manage a club as it is, even without doing marketing and outreach.
So I've taken up a project to try and help out. Bundle all the unusuals into one platform, help with club management, help with being seen. In a way, perhaps I can take all of these small activities and form "one big sport" that can grow faster as a result.
This is what I was looking for half a day yesterday! Will check it out. Sport is way nicer when social.
I was also curious what sports can help to replace specific gym exercises like dead lifting, only thought of cycling
TL;DR: highlight tailwind / bootstrap classes in VS Code.
Web Dev is only a distant area of expertise of mine. I find learning new things here not complicated, but a bit tedious.
What I'd like is to be able to scan through html files faster. And what hinders me the most, in terms of reading others' code now, is figuring which html classes are custom-declared, and which belong to a design framework, like bootstrap or tailwind.
Being able to tell at skim-speed would be awesome.
I'd find it nice to have a VS Code plugin to hide assertions. Specifically in Python, I see myself writing asserts to check for data boundaries, numpy array sizes, data types.
It takes space and diatracts a bit.
Bonus points if I could have a, for example "# /assertion_func" comment that would label a whole function to be folded together with asserts.
A lot of the music I like is on YouTube, including many obscure remixes and covers. Now, I'm not a fan of forgetting things and YouTube can be pretty volatile, with videos getting removed.
Rather than downloading all the audio, I made a playlist scraper for the names. Runs once a day on PythonAnywhwre, collects the names into a database, helps me sleep at night.
You can schedule "takeouts" from takeout.google.com to run every 2 months for a year - but as far as I can tell they NEVER fire off more than the initial time. Very weird yet so very Google too.
My mom's a textile designer, contracting with large companies, but somehow not getting to use any fancy tools they. She says design software existd, but it's pretty expensive.
At the core, she's making plaids - a black-and-white pattern (2d grid) corresponding to vertical and horizontal threads being above one another, and color schemes (2 arrays) for the vertical / horizontal strings (in most cases, does not match the size of the grid).
So she made her own monstrocity of a pipeline, with Photoshop, Plaid Maker (a rather old web app, https://www.plaidmaker.com/), screenshots, Excel (for designing stuff - yes!), and an old Soviet-era MS-DOS designer program (in Russian, of course).
Throughout uni, I was jokingly saying I'd build a better Plaid Maker for her, even if the web is not my area of work. After The Great Resignation I figured, why not?
Note: not for mobile, but should work on a tablet.
Privacy: no analytics are used, but it does rely on Google Firebase for authentication and auto-saving of your designs. Log in to keep them permanently; otherwise, they're saved on an anonymous account, cleared up once a week or so
I love this. I don't know much about plaids but I really really really enjoy hearing stories of people replacing DIY, hacky, inefficient or simply long and painful pipelines and workflows with custom well designed software.
It's always nice to hear how much it helped people and removed the problems the previous version had - bonus points if it's for a small business or a single person rather than some giant enterprise org. I hope one day to actually be able to produce software that can benefit people that way. All I seem to do recently is write software that gets killed off because politics all while dealing with poor quality code from other teams and it's really demoralising.
I would love to know if anyone has a site or something where these stories are documented!
I'm building a system that reads Slack, listens to Google Meetings, user complaints, etc and gives me prompts I could feed into coding agents or planners.
Problem-to-prompt seems like a larger obstacle than coding these days, I wonder if it's solvable, and if solving it makes cheaper coding agents viable.