I've been working on my own inference engine (rainference) to see just how hard I can push the AMD Radeon 9070 XT with qwen 3.8 27b
At this point I'm now more than double llamacpp at 84.8tok/s decode (this is with the DE sharing the card)
field report here: localmaxxing.com/en/models/Q…
@LottoLabs plz add rainference to the engineName enum so I can post speedtests :D
rain retweeted
"GLM is a danger to us all!"
GLM:
GLM-5.3 has helped defend 389 open-source projects, with 4,249 potential vulnerabilities found so far.
OpenVuln is still running. The service remains free, and findings go privately to maintainers.
huggingface.co/spaces/zai-or…
rain retweeted
Underrated accounts in the Local AI space to follow and keep an eye on
- @BrandonMusicKy
- @loktar00
- @mr_r0b0t
- @usr_bin_roygbiv
- @TechMDAI
- @net_termina
- @itsjustmarky
- @u1tra_instinct
- @QuixiAI
- @pupposandro
Opensource / Local AI is nothing without its people 💙
rain retweeted
The time is now.
I know it may feel like the harness game or the artificial assistant space is already too competitive.
I actually think none of these spaces have found the "one product to rule them all." I don't even think Anthropic and OpenAI are comfortable in their market positions.
I know there are a lot of big valuation numbers being thrown around, but the key thing to remember when you're this early in a cycle—which we are—is that the exact same product that failed last year could succeed this year.
The biggest advantage small builders have is patience.
I know a lot of builders follow me, and I imagine most of you have four or five different projects you're tweaking and testing every now and then.
The key is to never stop testing them, never stop pivoting them, and never stop adapting as a builder.
The things that were true when you first built them may no longer be true now—even if it's only been four months. Set aside a screen-free day once a month to reflect on this.
Big companies that raise large amounts of money have to show results for that capital. VCs are not the most patient people.
If you're a small builder and you haven't raised money, you have two huge advantages:
1. You can afford to be patient, wait for the right time to strike, and continuously improve your product as the underlying technology changes.
2. Nobody is rushing you to raise more or find their exit liquidity.
Take the time to do your craft. Building a beautiful, winning product is a craft. It takes care—it cannot be rushed.
There are people reading this right now who are going to be billionaires. And they're going to be billionaires because they internalize this message:
Think big.
You have so many more advantages than other people. The only thing holding you back right now is if you think too small—if you build a tiny little piece of the puzzle instead of building the whole damn thing.
Now is the time to think big.
To really think about how you can change entire domains, transform human behavior, and become someone’s favorite every day product.
I know thinking this big can be scary. If you show people your product, they might ask:
"Who are you to think you can build the ultimate assistant, the ultimate operating system, or a completely new way to interact with your computer?"
Who are you? You're the builder.
Like in the Middle Ages, you're one of the builders who is going to construct the great cathedral of our future. You may only be responsible for carving a single granite block to fit among all the others, but you are a builder. You are part of the collective building toward a future that looks completely different
This isn't about achieving transhumanism or some mythical singularity.
Because the more things change, the more they stay the same. This is just another cycle—an exponential jump in what we can do using technology.
You are at the very forefront, even if you don't realize it because you think, "Well, I haven't raised money... it's just me at a computer with a few tokens. There's companies raising millions..."
You really can build anything. You really can learn anything.
Never lose sight of the massive advantage you have.
The time to build is now.
this. it’s beyond taboo in gamedev to admit it.
everyone good in games is using ai and not saying anything because they have to hide from the pitchforks of their coworkers
I share this experience and pretty much agree
I can't understand how professional software developers are so attached to days of manually coding.
Was objectively a top-tier software engineer (e.g. hired as a quant dev to build high performance systems) and honestly not having to type another line of code feels like such a good riddance.
People being so forlorn about agents being good at coding is insane to me. You can continue to ride horses as a hobby, no one can take that away from you, but you should be happy for everyone else to have engines.
Coding has and should always (just) be a means to an end. The product was and always *should* be the point. It's undeniable that agents are just objectively better at getting to the point now.
Go ahead and manually write code as a hobby.
rain retweeted
For engineers having trouble with @dhh’s talk here some obvious truths about the future of programming:
You will not write code
You will not read code
You will not choose a language
You will design the architecture
You will define outcomes
You will build auto-delivery loops
rain retweeted
It’s crazy how many young Americans are so desperate for radical change but dismissive of the One Likely Shot We Have Right Now at facilitating it.
Apple is lost because bean counters and logistics men can't navigate paradigm shifts. Happy they put a hardware man back in the hot seat, but it's the software that needs fixing.
rain retweeted
dhh redemption arc
In his Rails World 2026 keynote @DHH just said that Basecamp is not using Ruby on Rails for the new Hey version. Replaced the Ruby on Rails backend with Rust. I totally agree with his argumentation. But this must be a shock for many developers. Hell just froze over.
here's how i shipped 2,500 PRs last month to production
this was originally supposed to be for Cursor Compile in London. i couldn't make it since i was livestreaming for Grok @Bot Galaxy so i'm making it available for free here on X! watch it on 2x speed, i talk slowly
This video is larger than Cloudflare's 512 MB cache, so it can't be played through. More donations are needed to cover a larger cache. Donate
We have had rsi since at least may 2025 arxiv.org/abs/2505.22954
Just a reminder that so-called "Recursive Self Improvement" is not a thing that's actually happening with AI, nor is a thing that can happen with statistical inference
Recursive self improvement is something that only a thinking machine can do and AI machines (linear algebra equation solvers) don't and can't think