• neural network is trained with deep Q-learning in its own training environment
• controls the game with twinject

demonstration video of the neural network playing Touhou (Imperishable Night):

it actually makes progress up to the stage boss which is fairly impressive. it performs okay in its training environment but performs poorly in an existing bullet hell game and makes a lot of mistakes.

let me know your thoughts and any questions you have!

repo: https://github.com/damienallen/urban-heat

cross-posted from: https://programming.dev/post/14939898

I wanted to share a small project I've been working on. The goal was to make the data from NASA's Landsat Thermal Infrared Sensor more accessible to the general public.

I worked with the raw temperature band data to general annual maximum surface temperature raster images for large urban areas covered by the Eurostat GISCO Urban Audit. In the browser, these images are transformed into easier to interpret isotherm contours with some adjustable settings.

I don't have a specific target audience in mind. The map could help identify areas of refuge for the warmer months, or overheated neighborhoods to avoid as we march towards a toasty future.

Feedback is welcome :)

“The most intelligent AI code generation tool out there and we have the data to prove it.”

been using for a few weeks; it’s quite competitive to copilot only coming shy of supermaven in speed and performance

https://codeium.com/blog/code-assistant-comparison-copilot-tabnine-ghostwriter-codeium

I am looking to program something similar to a simulation game, but free-form in its customization and scripting to the point where no strategy game will get me close enough.

I initially thought to start from scratch, simulating all the basics. Simulating money, people, resources, maps, etc. Obviously this is very ambitious.

Are there any libraries or frameworks that could help me with this? I don't want something opinionated that decides the model for how to simulate, for example, money or a person. I want to preserve the ability to simulate those with the models and math of my choosing. But maybe a library that has the foundations of simulation in general, so that I don't have to build everything completely from scratch?

I understand what I said sounds very vague. This will be something I will discover as I do more of it, so forgive the vagueness.

At the 2023 Worldwide Developers Conference (WWDC), Apple launched the highly anticipated new generation data management framework — SwiftData. As the successor to Core Data, can SwiftData play a key role in the Apple ecosystem? With WWDC 2024 approaching, this article will evaluate the overall performance of SwiftData since its initial release during the Xcode 15 period (i.e., its first major version), and provide a forecast of its future development.

JavaScript is a language that's easy to learn and has a gorillion features shoehorned in, to varying degrees of success. If you're relatively new to programming (just got done learning Python/Java) there's bound to be something new that you learn from JavaScript due to all its extra added features. Though these extra features are cool, over time you will slowly learn of the flaws that JavaScript has, and you will begin to associate "easy to learn with cool features" as the worst mistake that the language made. So, surely the answer must be to reduce how easy things are to learn. You will start to worship anything that excludes the less experienced and has cool features - for example Haskell or JS Frameworks - and you will denigrate solutions that seem too "simplistic and dirty". You will point to examples like C or Adobe Flash, citing that they have must have security vulnerabilities because they're accessible and so stupid people must be creating footguns with them. You will completely neglect the ways in which they're poorly designed and you will completely ignore the positives of both. Fundamentally, you will refuse to acknowledge that it's possible for software to be accessible and to be designed well at the same time.

Today, the internet is a dumpster precisely because of this false equivalency. It would be easy for Google to remove the guardrails from WebAssembly in some sort of public testing version of Chromium, allowing WASM to support fun little runtimes with 10x the safety of Flash. Artists could use software packages similar to the old Adobe Flash suite in order to make cool things again, expanding beyond the Neocities pages that are currently trendy. Over time, we could improve WASM environments to be incredibly safe, have interesting specialized runtimes and make super cool creations, a development model which basically already worked with HTML and CSS. But, because people still think accessible === vulnerable, we will never have that, and so every site will have the same hyperminimalistic slop look, and artists will be pushed onto the same shitty platforms to do nothing exciting in formats that have existed for centuries.

Programmers learned the wrong lesson in the 00s - that everything needs to be gatekept to protect people from themselves. The actual lesson was that designing things properly can let anyone make anything. Sure, the DOM doesn't satisfy people's need today, but it used to be excellently designed for its task - that's why it could let anyone build something amazing

Some background: I'm a software developer, and I've never really participated in the open source software community before. (i.e. I don't contribute to open source projects, I don't know anyone who does, and I don't really know anything about the companies who start these projects to begin with, or what their motivations are for being open source.)

I'm currently trying to find software that my team at work can use to solve a particular problem we have. After doing some googling, it looks like this open source product called OpenReplay is a good fit for what we need: https://openreplay.com/

But when I first visited that website, I noticed that the background artwork looks AI generated. This made me feel skeptical of the project, and it makes me wonder: what if it's actually a huge scam and it's actually malware? For example, maybe OpenReplay is actually a copy of a different legitimate product that I'm not aware of. Maybe all of the stars, forks, and discussions on the GitHub page are from fake accounts. When I Google OpenReplay, there aren't a whole lot of results. How do I know if it's trustworthy if I can't find an authoritative source telling me it is?

Maybe I'm just being paranoid. But this is basically the first time in my career where I've tried to vet a new piece of software for my team to use, and I want to make sure I'm doing it right. How do you know when a product like this can be trusted?

EDIT: I don't mean to cast doubt on OpenReplay specifically, I'm just using that as an example because it's the product I'm currently looking into. My question applies to any piece of software that isn't widely known about.

cross-posted from: https://lemmy.ml/post/15953505

SVT-AV1 2.1.0 just released, how does it compare to SVT-AV1 2.0.0? Well-known encoder Trix attempts to answer this question with metrics, graphs, and detailed analysis.

I just searched online and was taken aback by the lack of content I could find , there are millions of video for different small niche things by hundred of people which are right and wrong about things but the most I could even find about how chips are made today are the ones explaining how silicon works etc. LTT is the only one which even have a factory video and it is too very censored uninformative and useless for my questions .

1 - I get that light is flashed in binary to code chips but how does it actually fookin work ? What is the machine emmiting this light made up of ? How does this flashing light hold as data forever on chip ?

2 - How was program's, OSs, Kernal etc loaded on CPU in early days when there were no additional computers to feed it those like today ?

3 - I get internet is light storing information but how ? Fookin HOW ?

4 - How did it all come to be like it is today and ist it possible for one human to even learn how it all works or are we just limited one or two things ? Like cab we only know how to program or how to make hardware but not both or all ?

5 - Do we have to join Intel first or something to learn how most of the things work lol ? Cause the info available online about the software, hardware, skills etc is shit ? Not even RISC-V documentary are available .

Context - Just started learning python and got philosophical to how all things came to be ? Is just making apps or websites even a thing worth learning in the grand scheme of things ? I get that some people is just okay with that but come on have you never thought about how the deep you can go ?

anyway feel free to tell me to stfu and I'm sorry if sub=wrong and will move on request . And as the username suggest I'll be posting questions as I have them and thanks.

ALSO ELI5 everything please

When people reunite after a long absence, they are often surprised by the changes in each other; however, the transformations in those who are with us day after day are often overlooked. In this article, I will sift through the key updates to SwiftUI that have made a significant impression on me since its first version. This is not only a reflection on the evolution of SwiftUI from its inception to its maturity but also a fresh appreciation of the vitality it embodies.

Skip the beginning marketing copy. There's a lot of dev focused news

New favorite tool 😍