
VP, Product Experience @arriviatravel Ex Kickstarter, Amazon, Trulia. Entrepreneur. UX Designer & Developer. Husband, dad, snowboarder, mountain biker, gamer.
Seattle, WA
Joined April 2010
- Tweets1.6K
- Following2K
- Followers3.4K
- Likes322
Pinned Tweet
Designers need to validate their work in production just as much as engineers need to ensure their code works in production.
All too often designers seem to think if they do their homework (user interviews, prototyping, usability testing...) they have solved the problem.
11
The @nba needs to get their stuff together. What's being played now isn't basketball.
37
You're in a ride share (Uber, Lyft, etc), the driver is left lane camping and getting passed on the right. What do you do?
1
85
AI is great for design, but not in the way you might think.
It's great in that AI makes it so easy for so many people to build products. They will all have average or meh design. And with so many more products in the market, having good design will be even more of a differentiator than it's ever been.
87
Five days ago I told @claudeai I shouldn't have to make a user an editor on a doc for them to comment. Today I check and there's the option I want. I'd love to know how their process works to get it done so fast. 👏
70
One of my favorite things about AI, and maybe this isn't good, is now all of those moments where I only have five minutes free, there's a lot of tasks that can fit in.
Before it was like "I'd start this but I'd only gain context before I have to stop."
Now it's "I can start this, put it on the back burner and come back later to find something ready for the next step."
1
109
I don't use Windows a lot anymore. I just tried to add a printer. Spent an hour on it, still couldn't get it to work. Same printer took 30 seconds to set up on a Mac.
Even worse, I tried to delete a printer off the Windows machine and after 15 dialogs, running powershell as admin... I finally had to resort to deleting registry entries.
It's 2026, how is Windows still this bad?
1
103
Want a faster way to serve local Models on Macs?
MLX-Serve 26.10.1 is out: speed across the board.
Exactly faster. Qwen3.8 27B with its drafter
+66% on an M5 Ultra,
+28% on an M4 Max,
+37% on an M1 Pro.
Qwen Flash Next:
+11% decode on M4 Max,
+51% prefill on M5 Ultra. (up to ~5400 PP!!!!)
Gemma 4, Qwen, LFM2.5, Spark, Bonsai 2 all faster, output byte-identical to 26.9.6 (across 18 models).
New:
* GGUF models run on our own MLX engine, written in Zig and Metal, instead of the llama.cpp fallback. (experimental)
* 🍣 Sushi-format Flash Next packs run directly.
* Qwen-Image 2.1 edits pictures from instructions.
* Drafting is on by default for every model that supports it.
Huge thanks to everyone who filed, tested and fixed this one: @STRML_, @sbusso, @wutang_superfan , lborloz, LXD-8 @alinselea zeeshanhaque21 @Lojza3D, h9q2cyxvgm-ui @kennethrdegraff, @CowboyCoderHQ , codysk, @Beamsters1, @CerebralCoding_, @AjAbsaki , and more.. THANKS ! Please comment if I did not include your name, for visibility, I know people mostly by their GH handle.
Special thanks to @ashxhart and @Spangler3000 for pushing !
Benchmarked and tested across 5 machines: htmlhost.jax.workers.dev/ren…
GH Release Download + Changelog:
github.com/ddalcu/mlx-serve/…
Website: mlxserve.com
1
139
MLX-Serve is a private AI studio that runs entirely on your iPhone. Pretty sweet.
WHAT ? NO WAY !! This got approved in less than 10 hours !!
Thank you @Apple and @johnternus you are my new favorite CEO !!
Get it here: apps.apple.com/us/app/mlx-se…
And thanks to whoever put this on the reviewers' radar. Never happened in my ~19 years of developing for iOS.
This is going to sound crazy, but after ~30 years of web development, using CVS, SVN, TFS... I just pushed my first personal site using git.
Yes, I've used git for work for many years, but this is the first time I've used it for an individual project.
108
I asked Perplexity if I should cancel my @perplexity_ai account. It said I should.
I've used Perplexity for about 18 months. Over the last six months especially, the quality of responses has gone way down hill. Now rather than giving me an answer, it often responds with something like "I don't know, but if you prompt me again I can go search and find out." So I prompt it again and I get a similar response - it still doesn't bother to go get the answer. It's now nearly as bad as Copilot.
I hope they can turn things around, but the chat response doesn't seem hopeful.
1
1
130
I use Cursor way more than Claude Code. Tonight, I decided to use Claude for something. After a half hour of babysitting the execution of a plan, Claude got me banned from my own app.
1
2
152












