Paul Solt
@PaulSolt
Followers
3K
Following
9K
Media
2K
Statuses
16K
I love pour-over coffee, iOS dev, cooking, bread, running, gardening, and frisbee. Plant-based foodie! @BrewCoffeeApp Previously: GoPro, Apple, Microsoft
Rochester, NY
Joined May 2008
Who Codes Best? (GPT 5.1 vs Gemini 3 vs Grok 4.1) https://t.co/4pE8me0zZC
0
0
3
Each time an agent hands back work that’s not complete, you waste development cycles. Optimize your build tooling and you’ll be able to make huge progress with gpt 5.1 codex max. The agents smarter when it can spot the problem.
0
0
0
Ask your agent to use: https://t.co/1e6lEbfjd2 Too often an agent (codex or claude) will say your build is green, but it has failures it didn’t see. When you work with agents, give them tools to see the problem, don’t expect them to read walls of compiler output.
github.com
A little beautifier tool for xcodebuild. Contribute to cpisciotta/xcbeautify development by creating an account on GitHub.
0
0
0
If you try coding agents with Xcode iOS/macOS projects you need to ask your agent to create a Makefile or build script so that there’s zero guess work on how to build, test, and run. But more importantly use tooling to verify no build failures exist…
2
0
0
Apple Music collapsing navigation into one button is always wrong. I never want this experience. I want single taps to get to different playlists. So much of the iOS 26 redesign is backwards for usability.
0
0
0
Codex 5.1 Max Tried to Build My Mac App… Here’s What Actually Happened
0
0
0
Peek behind how Cursor trained their Composer model.
very honored to have @leerob grace the AIE stage for the first of hopefully many times. Cursor's execution is world-class and I heavily appreciate the detail Lee went into on Composer's arch and training - check out the custom MXFP8 kernels used, @anyscalecompute Ray load
0
0
2
Fizzy looks intentional and amazing! Reduce the clutter of a kanban style board and keep it fast and fluid. Nice work @jasonfried and @dhh !
INTRODUCING FIZZY Have you noticed that every issue and idea tracking tool you loved slowly morphed into boring, sluggish, corporate bloatware? Trello put on 40 pounds of cruft. Jira started charging by the migraine. Asana tried to become everything to everyone. GitHub Issues
0
0
2
I love how dedicated the Codex team is with fixing user feedback. Codex's CLI has come a long way... Initially it was barely usable, but now it unlocks a new paradigm of coding. There are still quirks, and I hope they can make Codex even better. @OpenAICodexCli @openai
Asked Codex for help on this one. After it scraped all the replies, there were over 260 unique suggestions, which it then categorized + prioritized and I re-shared with the team. Lots of these we're hoping to ship or improve before end of year. Thanks everyone!
0
0
1
I’m available to hire as an iOS dev! 👋 I’m open to remote, in-person, or hybrid. Either in the US or Europe. Looking for full time, something new and exciting I’m ready for a new adventure! Reposts appreciated and DM’s are open
7
60
172
@WisprFlow can you fix this Apple Pay issue? Another setting or different way of hosting the window/panel?
0
0
0
Wispr Flow breaks Apple Pay in Safari on Mac. Quit Wispr Flow to buy things. If you can't use Apple Pay with some Mac apps open, check and see if they have hidden windows stealing focus from Apple Pay checkout. https://t.co/2dZCpS4grH
1
0
2
codex 5.1 plans better before writing code than Opus 4.5. Always use the best model for the task... Or you'll end up with crashing bugs.
Opus 4.5 is still sloppy. Let codex review Clawd's fix and ofc it forgot a case because it doesn't read enough files upfront. To be successful with Opus you always need plan mode and all.. codex does that by default.
0
0
1
I gave Claude / GPT Codex / Gemini 3 (via Antigravity) a task to come up with refactor for a big refactor I wanted to do in my Swift project, I gave them pretty detailed expectations and goals I want to achieve with it then I scored the resulting plans: Codex is very close to
3
2
49
Amazing insights. I never thought about all the nuances of what goes into building an intelligence. It was a great view behind the scenes. Thanks @dwarkesh_sp and @ilyasut !
The @ilyasut episode 0:00:00 – Explaining model jaggedness 0:09:39 - Emotions and value functions 0:18:49 – What are we scaling? 0:25:13 – Why humans generalize better than models 0:35:45 – Straight-shotting superintelligence 0:46:47 – SSI’s model will learn from deployment
0
0
1
We’re looking for interns! I first worked with Kyle as an intern, and it was an incredible experience that led me to where I am today. I highly recommend you apply!
We're building the future of app development. And we want you on the team! Bitrig is hiring interns. Want a peek behind the curtain? Read about our Swift interpreter: https://t.co/79OIP2rvd8 Sound awesome? Learn how to apply here! https://t.co/0BhkiN06Es
1
1
2
housekeeping: https://t.co/GAz0nBJ6NL -> better trimming for comds-with-dash https://t.co/DLPpqiN4KX -> fixed an issue with OpenAI/Anthropic Status page parsing https://t.co/5fXnnnOpOn -> more aliases for clumsy agents and better handling of MCPs that need auth
github.com
Call MCPs via TypeScript, masquerading as simple TypeScript API. Or package them as cli. - steipete/mcporter
2
4
60
Amazing fun Mac app! Time to put up some lights!
This season’s Festivitas updates are here! 🎄🚀 Festivitas 1.9 turns your Mac into a winter wonderland with snow and adds Shortcuts automation ❄️💻 And for the first time, Festivitas comes to iPhone and iPad with animated festive lights on your Home Screen 🎄📱 Link in 🧵
0
0
2