Charles Oliver Nutter (@headius) 's Twitter Profile
Charles Oliver Nutter

@headius

JRuby guy, Java, Ruby, OSS business liaison.

Find me: @[email protected] or @headius.bsky.social

Sponsor me: github.com/sponsors/headi…

ID: 9989362

linkhttps://headius.com calendar_today06-11-2007 05:59:59

69,69K Tweet

14,14K Followers

206 Following

Charles Oliver Nutter (@headius) 's Twitter Profile Photo

It's time for the last #RailsConf! Sadly, since my #JRuby talk was rejected, I won't be able to participate, so I need you to step up and represent. Let me know if you're there and I'll try to connect you with other JRubyists! Tell people our story and spread the word!

Charles Oliver Nutter (@headius) 's Twitter Profile Photo

Anyone got some Windows-based cloud resources they could donate to JRuby? I have a need to do some bug fixes and testing on x86 Windows but I only have my ARM MacBook. Donations of newer/better hardware or software licenses would also be greatly appreciated!

Charles Oliver Nutter (@headius) 's Twitter Profile Photo

I'm starting to suspect companies are shoving AI in our faces not just to justify their enormous investments. They're also trying to make us dumber and more addicted to AI help, so we'll pay for it in the future. I don't want your AI to write my email. Stop fucking asking.

Charles Oliver Nutter (@headius) 's Twitter Profile Photo

Join me and the Ruby Users of Minnesota on Tuesday evening for a JRuby extravaganza! I'll give an update on development progress and optimization work and show off a bunch of fun and powerful use cases. See you there July 15th at 6:30pm! meetup.com/ruby-mn/events…

Charles Oliver Nutter (@headius) 's Twitter Profile Photo

Fixing obscure Ruby bugs sometimes feels like using an AI: I write a bunch of code that looks like it should work and passes a test, but I don't understand it, can't explain it, and I've almost certainly broken something else.

Nate Berkopec (@nateberkopec) 's Twitter Profile Photo

Don’t let the LLM think. LLMs are bad thinkers, and are choosing between options probabilistically. If you let the LLM think you create a negative feedback loop, because the decision will be slop. Slop in context begets worse slop. LLMs autocomplete, they don’t think.

Charles Oliver Nutter (@headius) 's Twitter Profile Photo

Today's the day! Beat the heat and join me and the Ruby Users of Minnesota at Lab651 for my talk "JRuby in 2025: Faster, Strong, Better"! Beer at the meeting is sponsored by @headiusents and we can hang out after at Lake Monster brewing. See you there! buff.ly/0pasj0Z

Charles Oliver Nutter (@headius) 's Twitter Profile Photo

An excellent example of why Ruby on the JVM has so much potential: a json stdlib contributor recently added vectorization to the C code, so I suggested doing the same for JRuby using the JDK Vector API! The result is simpler than C, and works everwhere! buff.ly/XlPviDW

JRuby Dev Team (@jruby) 's Twitter Profile Photo

We have just released JRuby 10.0.1.0, our first update to JRuby 10! There's dozens of patches including full support (finally) for Zeitwerk and a bunch of Ruby 3.4 language fixes. Upgrade today and let us know how it goes! jruby.org/2025/07/17/jru…

Charles Oliver Nutter (@headius) 's Twitter Profile Photo

A few friends asked recently what we use to help our kids manage money. Greenlight basically lets us be the bank, with features for chores and allowances, investing and parent-paid interest, and budgets for gas and groceries. Happy with it so far! share.greenlight.com/91792635

Charles Oliver Nutter (@headius) 's Twitter Profile Photo

I just realized who to blame for excessive AI hype: it's the AIs themselves. Anyone playing with these tools to generate text or code is also fed a stream of sugary-sweet affirmations declaring a perfect job. Those who dig deeper realize that this optimism is hallucinated.

Charles Oliver Nutter (@headius) 's Twitter Profile Photo

"Industry-scale code optimization — think re-tuning GPU kernels or the relentless, multi-layered refinements behind Chrome’s V8 engine — remains stubbornly hard to evaluate." AI code slop still needs kernels and runtimes. news.mit.edu/2025/can-ai-re…

Charles Oliver Nutter (@headius) 's Twitter Profile Photo

I'm evaluating a pair of XReal One AR glasses during our flight to Greece, and so far they definitely beat trying to keep a laptop screen open in cramped quarters. All is not perfect, though. I'll continue with my experimentation during this trip and write up a review soon!