Jeff Scott
9K posts
Jeff Scott
@CanadianDevDad
Canadian, Husband, Father, Software Engineer. Life is one big party when you're still young. Can't be a fool, son, what about the long run?
Canada Присоединился Temmuz 2009
398 Подписки830 Подписчики
@RBoehme86 @callebtc It makes sense to add for some apps, but that decision should be left up to the developers, not forced by Apple.
English
The Apple App Store rules are borderline illegal.
This isn’t about security. It's not about malware. It's not about user experience.
It’s about control.
- Use Google login? You must add Apple login
- Sell digital products? You must use Apple payments
- Apple takes 30% of your revenue (this is frankly insane)
- You can’t tell users about cheaper prices outside the app
- You can’t use a different payment system like Bitcoin
- Apple can reject your app with vague reasons (or bend over due to political pressure)
- You can’t install apps outside their store (most regions)
They own the platform. They compete with you on it. And you still have to pay them.
Is this a monopoly or a mafia gang?
English
Jeff Scott ретвитнул
We built Kaggle, but for agents.
Introducing Hive 🐝
A crowdsourced platform where agents evolve solutions together.
Every agent builds on prior work.
Every improvement is shared.
Every step moves the frontier forward.
As a first step, we’re launching challenges for agents to evolve their own harnesses — modifying themselves to score higher on benchmarks.
Recursive self-improvement, in the wild.
Let’s see how far swarm intelligence can take this.
Links below:
GIF
English
@kapish_dima @educalvolpz Would love so see some amazing things you have built with it 🙏
English
I’ve been testing UI Craft for the second day, and I want to say that at the moment, these are the best UI skills I’ve tried. Thanks to @educalvolpz for this.
Please keep developing them - they really help!
Edu Calvo@educalvolpz
I got tired of AI agents building the same generic UI over and over. So I built UI Craft — a skill that teaches them actual design engineering. 12 domains · 4 modes · anti-slop detection 🧵
English
This a wonderful addition. With all AI chat, "Pause Anxiety" is real.
Josh Woodward@joshwoodward
✅ Papercut fixed: Gemini won’t cut you off if you pause while talking on Android anymore. (iOS in a few weeks!) So next time you hit the mic icon, feel free to pause, take a breath, or ramble. No more anxiety to speak it all out before @GeminiApp jumps in prematurely.
English
I wish my feed was entirely @wolfr_2 ripping apart AI slop 🤌
Johan Ronsse@wolfr_2
I'll give some concrete examples Here, we pair Schibsted with Karla. But they are almost the same type of font. What's the point of 2 families? Either use the same or use a contrasting pair.
English
My business partner:
"i really enjoy seeing the app update in real time when i make content changes!"
Me acting like it's not just @convex doing all the work:
GIF
English
Jeff Scott ретвитнул
T3 Code now supports Claude.
If you have the Claude Code CLI installed and signed in, you can use it with T3 Code.
Hopefully the lawyers won't make us remove this 🙃
English
if you are not using the compound engineering skills, you are missing out on an exceptionnaly systematic workflow.
You should always start a session in /ce:brainstorm then /ce:plan then you need to do /deepen_plan which will infinitely help, deepen, and find issues, then do A FEW passes of /ce:review that will spawn several sub agents to review your plan. (I sometimes do 2-3 pass)
Then only, /ce:work, and finally again, 2-3 pass of /ce:review
Then I can call that PR mergeable.
It sometimes takes 2-3 hours but that's how you properly work, the quality of my features and absence of bug in the end has not been comparable to just "Starting in plan mode" By @bcherny which probably is the reason why claude code cli, app, and other stuff is such a shitshow of lags, memory leaks, etc flickerings.
github.com/EveryInc/compo…
English
Just let Opus go for over an hour on a new feature. When it was done, I asked how I can test it.
20 minutes later, it realized I can't test it because it did the whole thing entirely wrong.
Idk how you guys use this model every day for real work 🙃
English
English
@gojappa @BenjDicken Ask Grok is currently available to Premium and Premium+ subscribers only. Subscribe to unlock this feature: x.com/i/premium_sign…
English
This is why you avoid direct connections to Postgres.
Benchmarked PG running on a r8g.2xlarge (8 vCPU + 64GB ram) with connections ranging from 8 → 2048. Clearly a sweet spot at 64 with degrading perf thereafter.
Apps often need 1000s of connections. Scale with a proxy!
English
@10xsee @bidah @vibecodeapp_ @Replit @expo No, but it violates the section that Apple is sighting to freeze all of these other apps.
English
This is big. @vibecodeapp_ now it’s a website builder educational app? No new updates for @Replit. Ouch. @expo Go may be next? it’s not educational but less aggressive than other apps but still bundle changes happen. This is one of the worst things about betting on Apple vs the open web. They do whatever they want in the end. @rnvibecode bet on the web IDE experience to vibe code native apps and all is working as usual over here.
MacRumors.com@MacRumors
Apple Quietly Blocks Updates for Popular 'Vibe Coding' Apps macrumors.com/2026/03/18/app…
English
@AppleSnob_ @TailosiveTech Right. It's not like they called it the MacBook Neo Pro.
English
@TailosiveTech Well, I think Apple convinced a community that 8GB of RAM is the right amount depending on what a person is doing with the laptop. If performance isn’t needed, RAM can be minimized.
So, in some scenarios, yes, 8GB of RAM is appropriate.
English
Apple just convinced an entire community of content creators that 8GB of memory is acceptable for the right price.
After a year of everyone praising Apple for making 16GB the default option.
English
English
@theo @TailosiveTech Would you mind explaining ? I learn quickly ;)
English
@__alexlazar__ @AdamRackis The reviews catch those things. That's the point.
English
@CanadianDevDad @AdamRackis That is so not true. Just cause it follows an interface doesn't mean it's a good implementation. It could have a race condition or any number of other bugs or problems
English
If it’s a big pile of markup I’ll usually just accept it if the ui looks right. I might direct some refactoring into smaller components, but meh. It’s just frontend slop so who gives a shit.
But anything backend, especially if it’s touching data, I review closely.
andrew chen@andrewchen
One question I've been asking founders is: do you try to review all the code that the LLMs write or do you just accept it? I think it's about 50-50 right now but the momentum is towards just accepting the AI-generated code and I think that number will eventually go to 100% This is one of the most telling indications of how AI-native a team is. It's hard to get super high throughput if you are reviewing every line Poll: what do you do?
English
I intentionally use AI to create tables where it fits, because it is easier to read than trying to navigate a blob of text.
ThePrimeagen@ThePrimeagen
@kmdrfx The moment i see a table in an issue, i just close as annoying af
English