Ed “I’m Actually Mexican” White (@SpyHI) header
Ed “I’m Actually Mexican” White (@SpyHI) profile picture

🇺🇸🇲🇽 Señor Software Engineer. Army Veteran. 👨‍🎓 CS && Biz. ❤️ @DallasNagata. Organizer @JSConfHI. Tweets about tech, Hawaii & nerd stuff.

Honolulu, HI
Joined March 2009
Photo by Ed “I’m Actually Mexican” White (@SpyHI)Photo by Ed “I’m Actually Mexican” White (@SpyHI)Photo by Ed “I’m Actually Mexican” White (@SpyHI)Photo by Ed “I’m Actually Mexican” White (@SpyHI)Photo by Ed “I’m Actually Mexican” White (@SpyHI)Photo by Ed “I’m Actually Mexican” White (@SpyHI)Photo by Ed “I’m Actually Mexican” White (@SpyHI)Photo by Ed “I’m Actually Mexican” White (@SpyHI)Photo by Ed “I’m Actually Mexican” White (@SpyHI)Photo by Ed “I’m Actually Mexican” White (@SpyHI)
Ed “I’m Actually Mexican” White retweeted
vogue taiwan january 22 'heat wave' vs january 25 'overconsumption' both photographed by zhong lin
5
1,385
47
7,223
160,653
Giving a talk on AI/ML for Designers, a la @Wattenberger’s “hard and soft” essay, so not “how to use AI tools to help with design” but instead “how to reason about AI as a design solution.” I’m struggling with a title because “AI/ML in UX” is too vague. Ideas?
1
274
The idea of this talk is to give UX designers a sense of what AI engineers could reasonably implement, and the ecosystem beyond LLMs. Think “here are design opportunities posed by the web platform” except about ML. How can I get that across in a pithy title?
187
Ed “I’m Actually Mexican” White retweeted
The most annoying problem for Zero's launch was what bug tracker to use. GitHub is way too slow, and Linear doesn't have public bugs or permissions – both required for OSS projects. This was driving me nuts, until I realized the only possible solution… ⏩ (link in reply)
26
46
26
436
134,639
Me to @GenePark about finishing Mouthwashing last night
1
169
“We ought to only read books that bite and sting us. If the book we are reading doesn’t shake us awake like a blow to the skull, why bother reading it in the first place?” - Franz Kafka Mouthwashing lived up to the hype.
“Mouthwashing” is a message. Writer Johanna Kasurinen tells me: “I hope certain people in the gaming industry play this game. I hope they see themselves. And I hope it hurts. I hope it hurts, hurts, hurts.” My review on my 2nd best game of the year. washingtonpost.com/entertain…
222
Ed “I’m Actually Mexican” White retweeted
Microsoft just released a tool that lets you convert Office files to Markdown. Never thought I'd see the day. Google also added Markdown export to Google Docs a few months ago.
126
714
131
10,356
845,808
Ed “I’m Actually Mexican” White retweeted
I think programming GPUs is too hard. Part of the problem is sprawling, scattered documentation & best practices. Over the past few months, we’ve been working to solve that problem, putting together a “Rosetta Stone” GPU Glossary. And now it’s live! My take-aways in thread.
41
355
37
2,746
242,618
Ed “I’m Actually Mexican” White retweeted
“The state of most social platforms right now is that users are locked in and developers are locked out. We want to build something that makes sure users have the freedom to move and developers have the freedom to build.” — Jay Graber, CEO bsky.app/download
76
149
18
674
159,790
Ed “I’m Actually Mexican” White retweeted
reminder that the bcrypt hash function ignores input above a certain length! so if you do bcrypt(username || password) for some reason, a sufficiently long username will make it accept any password. to fix this you can sha256 the input first.
Okta allowing login bypass for any usernames with 52+ characters is insane Official Security Advisory: trust.okta.com/security-advi…
62
987
94
6,093
748,065
Ed “I’m Actually Mexican” White retweeted
"does bacon work with bun? when I try to run chai in fish my gerkin is complaining" This is what we sound like, friends
Replying to @joshmo_dev
does bacon work with bun? when I try to run chai in fish my gerkin is complaining
31
214
10
2,548
169,464
Ed “I’m Actually Mexican” White retweeted
Can y'all share your handles on Bluesky?
36
2
1
25
10,353
You know, I feel impostor syndrome as a an engineer *a lot* but then I put together full stack dev environment complete with a reverse proxy and hot reload using docker/vite+react/fastapi from scratch in a few hours and realize I have learned something after all.
1
5
139
Damn, honestly if Jane can be impacted, no one is safe 😱
Still trying to process this but my role at Meta has been impacted. Thank you to everyone for my wild journey at Meta If anyone is interested in working together on software/security, please reach out via my email / LinkedIn noted on my website 🫡
1
396
Trying to clear out my old phone browser tabs from stuff I meant to read, and stumbled on this article about how 80% of winning is not blundering, then listing a ton of classic startup blunders. Truly excellent! longform.asmartbear.com/avoi…
1
1
125
Finished @aarondfrancis’ High Performance SQLite to learn more about my favorite included-in-Python prototyping database. Enjoyed the concise, crisp videos and got some ideas about multi-tenancy and getting multiple writers that had never occurred to me! highperformancesqlite.com/
2
1
15
3,486
The good thing is that this doesn’t just apply to LLMs, this is probably foundational to design using any AI system—and to some degree, even managing humans. As vagueness and scope of a problem increases, the less likely it is a neural network of any type can solve it.
1
146
Re: scope management for “organic neural nets” success, I’m reminded about @SlexAxton’s talk about design systems and Robin Rendle’s essay on hyperobjects, which is worth a read in this context. “Division of cognitive labor” is just as important for AI. robinrendle.com/essays/syste…
1
1
138
That said, small scope, crisp division of labor, and bounded outcomes implies LLMs may not be that necessary for (optimally?) productive soft/hard systems. We know how to train those kinds of AI/ML systems already, just perhaps don’t know how to use them in system design…yet.
1
1
100