lej

2.2K posts

lej banner
lej

lej

@lej_dev

building things

NYC 加入时间 Ekim 2013
1K 关注6.2K 粉丝
lej
lej@lej_dev·
@cursor_ai please allow 2+ models to chat directly with each other about the best solution and compare their plans
English
0
0
0
54
lej
lej@lej_dev·
hey @Apple maybe don't un-download apps I've used in the last month
English
0
0
0
145
lej
lej@lej_dev·
@charliebcurran wouldn't call this "live action" but still, super cool
English
0
0
0
150
lej
lej@lej_dev·
@ErnestoSOFTWARE @maubaron took me less than 24 hours to get screen time approval last week, shouldn't take long!
English
0
0
0
95
Ernesto Lopez
Ernesto Lopez@ErnestoSOFTWARE·
we finished our app in 7 days it blocks your phone until you walk. But now, we are stuck.. WATING for the screen time apis to get approved by apple so In the mean time I decided to design some achievement badges for gamification and when i say desing, i mean we just sent a prompt to @rork app builder using gemini 3 and it generated them + added them to the project in 1 prompt. Meanwhile @maubaron was setting up the MMP to start running ads hopefully we get approved later today Or tommorow And we can finally submit Step Lock And scale this to $30k/mo
Ernesto Lopez@ErnestoSOFTWARE

We built this app is less than 7 days and it changed my life its an app that blocks your phone until you get your steps in its connected to real apple health data we built it using @rork with Claude Opus 4.5 and I literally use our own app every day you can see the chart in this video My steps per day went from unhealthy levels to above average we are beyond excited to launch monday or tuesday once apple approves the screen api always solve your own problems so you can be excited about showing up everyday.

English
21
5
116
17.8K
lej
lej@lej_dev·
@luigidemeo So it’s not local. You’re using the Claude API. And you can already do the coding task in the cloud with Cursor Cloud agents or Claude Code agents.
English
0
0
0
123
Luigi D'Onorio DeMeo
Luigi D'Onorio DeMeo@luigidemeo·
Clawdbot is the first time I really thought AI would change everything. Still early days for me, but one setup that's surprisingly useful: I have it running on a spare Mac Mini as my "background dev + life admin" agent, connected mainly via Telegram (quick from phone) and previously Slack. Cool example that's saved me real time lately: I set it up to monitor my Claude coding sessions and handle the loop autonomously when I'm away from keyboard. Workflow goes like this: - I kick off a project in Claude (e.g., building a quick tool for flight price scraping across providers). - When I hit a snag or want it to iterate (fix failing tests, add error handling, refactor), I just message it in Telegram: "Continue the flight-cli branch: tests are failing on provider X, try patching with retries + better logging. Commit when green." - Clawdbot pulls the repo (via its shell/browser tools), opens VS Code or whatever, runs the tests, sees the failures, thinks step-by-step (I set thinking to high), generates fixes, applies them, re-runs, commits/pushes if clean, and pings me back: "Fixed, pushed commit abc123. Tests passing. Next?" - If it needs more context (e.g., "Why did that API change?"), it browses the provider docs, extracts relevant bits, and includes them in the reply. Bonus: It also watches my calendar + traffic APIs in the background. If I'm heading to an IRL thing like kids sports practice or a meeting and traffic looks bad, it messages me 30 mins early: "Leave in ~15 min to make it on time – current ETA 28 min via Google Maps." It's not flawless, sometimes it over-thinks a simple fix or needs a nudge... but it's handling multi-step dev loops without me context-switching every 5 mins. Inbox side: it's quietly unsubscribing junk and drafting short replies on low-priority threads, so my Gmail isn't a warzone when I check. All local, no cloud leakage, costs basically zero beyond the Claude API I was already paying for. Setup was straightforward. If you're into self-hosted agents that actually execute instead of just chatting, give it a spin.🦞 clawd.bot / github.com/clawdbot/clawd…)
English
48
10
149
35.8K
lej
lej@lej_dev·
@zelmanski_tom Too bad no one wants to use that within a dating app
English
1
0
6
4K
Tomasz Zelmanski
Tomasz Zelmanski@zelmanski_tom·
yo, who did the Places app (from Raya)? really nice design
Tomasz Zelmanski tweet mediaTomasz Zelmanski tweet mediaTomasz Zelmanski tweet mediaTomasz Zelmanski tweet media
English
27
39
1.8K
169.1K
Collector Crypt
Collector Crypt@Collector_Crypt·
We’re dropping our highest value Gacha ever, @solana users, this is for you 💎 $1000 per pack, 2% positive EV target, 93% buyback rates 🎴Nothing compares Full of grails, including: 👻 PSA 10 Skyridge Gengar ($40,000) 🌊 PSA 10 Aquapolis Lugia ($35,700) 🔥 PSA 10 Shining Charizard 1st Edition ($23,000) With ~50 cards over $10k, @Collector_Crypt’s liquidity is deep ⛓️ See you there trainer, link below 👇
English
23
18
153
23.2K
lej 已转推
Peter Yang
Peter Yang@petergyang·
We're rapidly entering a time where building is more fun than consuming.
English
270
180
2.2K
80.6K
lej
lej@lej_dev·
@binji_x @instagram You can use hashes and cryptography without “crypto”. This truly does not need to be on a blockchain
English
0
0
0
57
lej
lej@lej_dev·
@lydiamoynihan This is like saying people lobbying against oil should shut up and just drive an electric car. Single person action won’t do anything. The system has to change.
English
0
0
0
54
lej
lej@lej_dev·
@amix3k Totally disagree, the whole reason I use cursor is their constantly improving UX for dealing with the constantly changing world of AI coding. Fewer experiments? No thanks!
English
1
0
2
311
Amir Salihefendić
Amir Salihefendić@amix3k·
Cursor feels like a company without product management. Just devs and designers running seemingly random experiments with little higher-level strategy. I use Cursor a lot, and it is becoming a very frustrating experience: — The interface is cluttered with tiny icons that get shuffled around every week — Are they building an agent or an editor? Pick one. This hybrid approach, where they pivot weekly, is frustrating — “Everything and the kitchen sink” isn't a great way to manage a product. A better approach is focusing on the core, and make that great instead of changing/adding useless things — Focus on stability. People use this for real work. Running into bugs and issues daily isn't the experience anyone wants Compare this to Claude Code: one text box, clear identity as an agent. Simple and focused. It's no wonder it's growing a lot.
kyzo@ky__zo

guys at @cursor_ai can you please slow down with changing the ui every 2 days? i'm like constantly confused between the agent and editor, feels like the buttons change every day. just make a decision and stick with it becaus re-learning ui every day suuuucks

English
169
39
1.6K
295.7K
lej
lej@lej_dev·
@vivarantex @ItakGol so much AI cope. will never understand why humans fear change so much
English
1
0
1
16
Vivarantex
Vivarantex@vivarantex·
@lej_dev @ItakGol same I told him haha he must be a noob at this and it's totally underestimating AI. Most people in the comments are just coping
English
1
0
1
21
Itamar Golan 🤓
Itamar Golan 🤓@ItakGol·
What it's like to watch AI fix a bug
English
435
2.6K
20.9K
880.5K
Interesting AF
Interesting AF@interesting_aIl·
A trick to remove something in Photoshop
English
73
1.4K
16.4K
1.5M
Jeff Tang
Jeff Tang@jefftangx·
Hot Take: Anthropic will acquire Linear. Coding is solved post-Opus 4.5. All the devs building CLIs and IDEs are working at the wrong level of abstraction. Why? The current bottlenecks in Agents are: - PR Reviews (h/t @rfgarcia). For side projects, Greptile can autonomously handle PRs (@Steve_Yegge didn't write a single line of code for Beads, 140k LOC). For prod deployments, the CTO still needs to have an understanding of what is happening, and to socialize those learnings organization-wide for cross-functional product roadmapping. - Agent Orchestration - This is being solved by Agent Cloud Mail by @doodlestein (built on Beads) , @TaskmasterAI by @EyalToledano. @intelligenceco is trying to solve this in the general agent scenario, but it will be solved in coding first. - Memory, but not an issue in coding with task decomposition and graph-based dependency planning a la Beads. The future will look like Technical PMs/Designers, CEOs, CTOs, and Staff Engineers working out of Linear/Asana (the company coordination layer), directing Swarms of Coding Agents, occasionally dropping into Cursor/CC for deeper understanding and debugging. Generation is solved. The frontier is Orchestration - the logical next step on Karpathy's "autonomy slider". 3-6 months.
Jeff Tang tweet media
Peter Gostev (Visiting SF)@petergostev

OpenAI or Anthropic will acquire @linear, let's see who gets there first

English
92
113
2.1K
372.5K
lej
lej@lej_dev·
@thesamparr Hire great people so you can work lesss
English
0
0
0
51
Sam Parr
Sam Parr@thesamparr·
work getting jacked hanging out w/ the fam no idea how there's time for anything else!
English
93
33
963
55.2K
lej
lej@lej_dev·
@kiwiasfi Purpose in your work
English
0
0
0
14
KIW
KIW@kiwiasfi·
without saying drugs... what is the cure for depression?
English
11K
352
6.7K
3.1M
lej
lej@lej_dev·
@_avdept @updatifyio Well most apps that compete with incumbents do so bc they have at least one killer feature the big guys don't. what is that for your app?
English
0
0
0
24
Alex 🔔 | updatify.io
Alex 🔔 | updatify.io@_avdept·
@lej_dev @updatifyio fair enough, I'm working on more features same story with pretty much any other new products the only difference - I noticed most other apps arent updating anymore, their changelogs stuck in 2023-2024 while mine actively developed
English
1
0
0
249
Alex 🔔 | updatify.io
Alex 🔔 | updatify.io@_avdept·
It’s been a month since I got my last paid customer on @updatifyio No matter what I do - no results I feel like I miss some basics What are some good guides how to get started in saas marketing ?
Alex 🔔 | updatify.io tweet media
English
130
2
183
613.9K