deepj retweeted
"Incoming transmission!" 📡
Classic Command and Conquer games are now open source—ready for modding, preservation, and innovation!
Read more: ea.com/games/command-and-con…
Dive into the code: github.com/electronicarts
deepj retweeted
Who remembers this moment in Prince of Persia (1989) where a magic mirror suddenly blocks the hero's path (and what happens afterward)?
This moment was actually never planned to be in the game. My original vision was a game with lots of traps, but NO enemies... (1/10)
deepj retweeted
These 94 lines of code are everything that is needed to train a neural network. Everything else is just efficiency.
This is my earlier project Micrograd. It implements a scalar-valued auto-grad engine. You start with some numbers at the leafs (usually the input data and the neural network parameters), build up a computational graph with operations like + and * that mix them, and the graph ends with a single value at the very end (the loss). You then go backwards through the graph applying chain rule at each node to calculate the gradients. The gradients tell you how to nudge your parameters to decrease the loss (and hence improve your network).
Sometimes when things get too complicated, I come back to this code and just breathe a little. But ok ok you also do have to know what the computational graph should be (e.g. MLP -> Transformer), what the loss function should be (e.g. autoregressive/diffusion), how to best use the gradients for a parameter update (e.g. SGD -> AdamW) etc etc. But it is the core of what is mostly happening.
The 1986 paper from Rumelhart, Hinton, Williams that popularized and used this algorithm (backpropagation) for training neural nets:
cs.toronto.edu/~hinton/absps…
micrograd on Github: github.com/karpathy/microgra…
and my (now somewhat old) YouTube video where I very slowly build and explain:
youtube.com/watch?v=VMj-3S1t…
deepj retweeted
Introducing 𝐌𝐨𝐛𝐢𝐥𝐞 𝐀𝐋𝐎𝐇𝐀🏄 -- Hardware!
A low-cost, open-source, mobile manipulator.
One of the most high-effort projects in my past 5yrs! Not possible without co-lead @zipengfu and @chelseabfinn.
At the end, what's better than cooking yourself a meal with the 🤖🧑🍳
deepj retweeted
You can now access the prism Ruby parser in JavaScript (node or the browser) through the @ruby/prism npm package.
This gives you access to all of the rich AST information that you can normally access through Ruby/C/Rust, but now in the browser!
github.com/ruby/prism/blob/m…
deepj retweeted
How we run projects at @linear:
1. Engineer is often the project lead. We rotate the role
2. Project lead gets things going and sends the updates
3. Write project updates after project meeting for the take-aways or decisions which everyone in the company can see and read
4. Debates around product decisions, not technical side is usually sorted out easily
5. Focus more on milestones than the project timeline. Each project has their own set of milestones
6. Internal “Feature roasts” to get live feedback about the feature
We are continuing our “How Linear uses Linear” series with a deep dive on how we run projects:
› Writing project specs
› How we use milestones
› Keeping others up to date on project progress
↓
linear.app/blog/how-we-run-p…
deepj retweeted
At @linear we optimize for solving customers problems & craft – not A/B tests, internal frameworks/OKRs.
To optimize for craft, give the builders control. To solve customer problems well, try to make the whole team have customer understanding.
Then combine these for the magic
deepj retweeted
🍄 20 years of Amanita Design! It's hard to wrap our own heads around it, but it's really been two decades since our first Samorost game ventured into the world. Well.. juchu! 🥳
To celebrate, we're throwing a massive 2-week-long sale on Steam, with our biggest discounts ever - up to 85% off! You can also find similar deals on the iOS App Store and Google Play. And, don't miss the special anniversary artwork featuring many of your favorite Amanita characters, made for the occasion by Jan Chlup of the Creaks team 🎨
deepj retweeted
Replying to @nareni0
Here are some people in the community talking about other things than Rails.
Not all of them are on this platform but will try to tag some people
- @schwad_rb is working on Scarpe that is a continuation of shoes and will make creating desktop apps so easy
- @duckiedevshow did a long series of streaming building a game with DragonRuby
- @joeldrapper talks about Phlex and recently he is building a lot of capabilities around literals and types
- @kowfm is refreshing campfire - an old but still fresh small web framework
- @HanamiMastery and @sebwilgosz are talking about Hanami 2
- @maciejmensfeld is talking about Karafka
- @marcoroth_ is doing a lot of work on Stimulus and now he is building gem.sh
- @ioquatix is talking about a new test runner - sus
- @jeremyevans0 is continuously evolving Roda
- @jankomarohnic is talking about rodauth and alternative to devise
- @julian_rubisch is posting about Hotwire and Stimulus
- @joemasilotti is talking about Turbo Native and Hotwire
- @_williamkennedy is talking about Turbo Android
And probably there are many more. These are just recent ones that I read in the last days that I remember now. Some of them are posting on other platforms so you might not see all of them here.
deepj retweeted
Artist Tang Zixi street drawings take advantange of cracks, wall imperfections and even weed: they are improvised on location through a process known as “pareidolic anamorphosis”
[more clips: buff.ly/44MgT6r]
xitter.cf/jin_nuko2/status/15604…
deepj retweeted
I wrote about how you can implement Ruby pattern matching on your Ruby classes at fly.io/ruby-dispatch/pattern…
Its such a great way to cleanup Ruby code that checks conditions on complex data structures like access control checks, HTTP routers, API responses, etc.
deepj retweeted
For many, @getajobmike is the embodiment of the commercial open source dream. He is in it, and he is in it for the right reasons, which makes @sidekiq and @contribsys super successful.
This little clip from my interview with Mike is the reason I strongly believe in #coss
deepj retweeted
I'm convinced everyone who actually likes JS/TS and the whole ecosystem is just suffering from Stockholm syndrome paired with being forced to use it. We're all just stuck with this reality. 😵💫 Layers and layers of madness, pure madness.
deepj retweeted
I wrote an article explaining how to manage Kafka offsets alongside SQL DB operations in a transactional fashion using Karafka:
mensfeld.pl/2023/06/inside-k…
A must-read for #ruby and #rails devs aiming to master event-driven architecture and #apachekafka
deepj retweeted
It's worth watching Sal Khan's TED talk for a preview of what Khan Academy have been working on using GPT-4 - building tools for students and for teachers ted.com/talks/sal_khan_the_a…
deepj retweeted
What is ChatGPT doing ... and why does it work? From the lore of neural nets to what Aristotle didn't get to ... here's my version of the story: writings.stephenwolfram.com/…
deepj retweeted
Art made by nature.. ❄️ xitter.cf/buitengebieden/status/…
deepj retweeted
Ruby 3.2.0 Released ift.tt/9lZHT5p
deepj retweeted
In memory of @ChrisGSeaton, a giant. tenderlovemaking.com/2022/12…





























