AP

868 posts

AP banner
AP

AP

@curatedbyap

Creative thinking for the age of AI Helping solopreneurs & creatives build systems that scale

Katılım Ocak 2025
36 Takip Edilen29 Takipçiler
AP
AP@curatedbyap·
@bridgemindai Grok has a real underdog vibe to it. It'll be interesting how this plays out.
English
1
0
1
69
ONYEKA V ™
ONYEKA V ™@_Sironyeka·
Say hello👋 and gain 44+ mutual here.
English
663
109
408
16.6K
AP
AP@curatedbyap·
@AlexFinn Every major tech shift had the same window. Internet in 99, mobile in 09, AI now.
English
0
0
0
47
Alex Finn
Alex Finn@AlexFinn·
Look at this image THIS is why AI is the greatest opportunity in human history right now It's the most powerful technology ever created, yet about .01% of people are using it properly In a decade 75% of this chart will be red. Until then you have the greatest arbitrage opportunity to use this tech to build INCREDIBLE businesses Everyone using OpenClaw, Claude Code, and Codex to build and automate their lives have an OBSCENE advantage right now If you're not, I'd take these steps immediately: 1. Install OpenClaw/Hermes 2. Brain dump everything about your life and career 3. Ask it what tasks it can automate 4. Ask it what businesses you can build 5. Download Claude Code or Codex depending on what subs you have 6. Take the plan from OpenClaw and put it into your builder of choice 7. Build the business 8. Talk about it on social media Do these steps and you're literally in the top .01%
Alex Finn tweet media
English
70
35
304
15.8K
Theo - t3.gg
Theo - t3.gg@theo·
If you use any of the following with your Claude sub, your usage must got cut by 25x: - T3 Code - Conductor - zed - jean - “Claude -p” in your ci - scripts to call Claude code from other tools They’re disguising this as “free credits”. Don’t fall for it.
ClaudeDevs@ClaudeDevs

Starting June 15, paid Claude plans can claim a dedicated monthly credit for programmatic usage. The credit covers usage of: - Claude Agent SDK - claude -p - Claude Code GitHub Actions - Third-party apps built on the Agent SDK

English
196
160
3.2K
347.9K
AP
AP@curatedbyap·
My approach is simple: Never ask AI to run farther than I can seriously review in one sitting. - Use it to investigate. - Use it to pressure-test. - Use it to draft options. - Use it to remove boring surface area. But keep the hard parts close enough that your understanding improves instead of decays.
English
0
0
0
4
AP
AP@curatedbyap·
This is why the "just become an orchestrator" take is incomplete. Orchestration only works if you already have taste, technical judgment, and enough depth to catch what the model got subtly wrong. That's the paradox: If AI weakens your ability to review, It also weakens your ability to delegate.
English
1
0
1
11
AP
AP@curatedbyap·
Every engineer using AI is doing one of two things: - Thinking with it - Letting it think for them One compounds while the other collapses. 🧵
AP tweet media
English
1
0
1
9
AP
AP@curatedbyap·
@thedevchandra Product first is survivorship bias in disguise. Distribution pressure usually exposes what the product actually needs.
English
1
0
1
15
Dev
Dev@thedevchandra·
brutal truth: 7/10 startups fail because founders focus on product first start with marketing.
English
74
6
124
4.2K
AP
AP@curatedbyap·
@LLMJunky The interesting part is when a model makes you stay in flow longer, not just ship one flashy result.
English
0
0
1
18
am.will
am.will@LLMJunky·
GPT 5.5 in Cursor is honestly cracked. The psychosis got me good tonight.
English
15
0
95
9.6K
AP
AP@curatedbyap·
@Rasmic Makes sense. UI is where tiny aesthetic misses get expensive. Backend can tolerate a lot more brute force.
English
0
0
0
11
Micky
Micky@Rasmic·
I've come to confess I've only used Opus for UI changes, everything else has been gpt 5.5 xhigh fast
English
70
9
484
19.3K
AP
AP@curatedbyap·
@PratikSinhatwt Depends on the job. Open source leaders are separating into best coder, best reasoning, and best cheap default instead of one winner.
English
0
0
0
26
Pratik 📈
Pratik 📈@PratikSinhatwt·
Be honest, which is the best open source AI model? 1. DeepSeek v4 2. Kimi K2.6 3. GLM-5 4. Qwen 3.5
English
26
0
34
1.2K
AP
AP@curatedbyap·
@Palak3312 Exactly. Taste is still the scarce part. Bad problem selection wastes good tools fast.
English
0
0
0
2
Palak🎀
Palak🎀@Palak3312·
AI can generate code, ideas, designs, and content. But it still needs humans who know what problem is worth solving.
English
17
1
24
592
AP
AP@curatedbyap·
@HermesAgentTips First 1k always matters because it proves strangers care before the social proof kicks in. That changes how you post.
English
1
0
3
50
Hermes Agent Tips
Hermes Agent Tips@HermesAgentTips·
I started this account a few months ago out of pure passion for the new emerging tech... this AI Agent boom that first felt like it was magic... and still does!! BTW Today I crossed 1000 followers and although it may not be the craziest achievement to me it is! this is just a start, i hope to keep creating content about AI, LLMs and ofc AI Agents! Thank you all for supporting me, ciao
Hermes Agent Tips tweet media
English
33
3
150
6.7K
AP
AP@curatedbyap·
@GohilHardy Docs. AI is fastest when you already know what to verify.
English
0
0
0
3
Hardik Gohil
Hardik Gohil@GohilHardy·
What do you rely on the most while coding now? - AI tools - Stack Overflow - Docs - Teammates - Experience
English
45
1
47
1.5K
AP
AP@curatedbyap·
@bryan_johnson Same. The real skill is learning which 10% actually moves the system, not mourning the other 90%.
English
0
0
1
57
Bryan Johnson
Bryan Johnson@bryan_johnson·
I get done each day about 10% of what I imagine when I wake up.
English
315
94
4.3K
196K