Carlos Marchan (@CarlosMarchanJr) header
Carlos Marchan (@CarlosMarchanJr) profile picture

AI Agentic Engineer · 0→1 & 1→N

Miami, FL
Joined August 2020
Photo by Carlos Marchan (@CarlosMarchanJr)Photo by Carlos Marchan (@CarlosMarchanJr)Photo by Carlos Marchan (@CarlosMarchanJr)Photo by Carlos Marchan (@CarlosMarchanJr)Photo by Carlos Marchan (@CarlosMarchanJr)Photo by Carlos Marchan (@CarlosMarchanJr)Photo by Carlos Marchan (@CarlosMarchanJr)Photo by Carlos Marchan (@CarlosMarchanJr)Photo by Carlos Marchan (@CarlosMarchanJr)Photo by Carlos Marchan (@CarlosMarchanJr)
This is my submission for the @NanoBanana weekend Hackathon. Going through an expansion in my career I put together a job hunt tool to help me with the pain of applying ... in the making 😉
1
284
Carlos Marchan retweeted
Prompt of the day is a simple one: "How do we make this change fully verifiable?" Makes the agent bend over backwards to verify all its changes. Signing up for real services, clicking through the frontend. Quality goes through the roof.
242
112
29
2,841
130,676
Jev is AI on steroids
Jev might genuinely be an “Internet moment” for AI. TypeSafe reports up to 193x faster and 444x cheaper performance in tests with Claude Fable 5.1 and GPT-6 Astra. @0xCodila just wrote a great 10-page article explaining what Jev is, how to use it, and where its 100x advantage comes from. Here are the 10 steps: 1 → LLMs create. Agents act. Jev decides the next move. 2 → Turn agent forks into three primitives: Choice, Score and probability. 3 → Build with OpenAI, Anthropic or xAI first, then swap Jev in without rebuilding the graph. 4 → Start with shared state, parallel decisions, risk thresholds and an execution queue. 5 → Batch decisions instead of making them sequentially. In one test, 13 questions were 10x faster and 12.2x cheaper. 6 → Put Jev at bounded forks: agent, model, tool, browser action or human escalation. 7 → Benchmark the whole loop, not just individual model calls. 8 → Rank wide, read narrow: shortlist first, then spend compute on what matters. 9 → Reuse the same system: State → Questions → Action → Verify. 10 → Keep Jev out of math, writing and irreversible execution. Code computes, LLMs create, Jev decides. The result: A slow, expensive agent loop becomes a much faster decision system that can route, score and escalate in milliseconds. Full breakdown below ↓
88
👀
Replying to @trq212
AGENTS.md support is built off of Claude Code mods, our upcoming way to customize the Claude Code harness. This is a built-in mod, but you’ll be able to build custom versions of project instructions yourself as you’d like too.
82
Carlos Marchan retweeted
6
1
32
1,450
Carlos Marchan retweeted
Anthropic just released a FREE 4-hour course on using Claude like its own engineers. 🤖 If you use Claude for coding, research, or productivity, this is worth watching. What you'll learn: ⏱️ 00:15 — Prompt Claude the right way ⏱️ 33:21 — Why Claude struggles with code ⏱️ 01:33:39 — Anthropic's daily workflow ⏱️ 02:50:56 — The tweak that makes Claude much smarter Whether you're a beginner or an AI builder, this course offers practical techniques you can apply immediately. 📌 Bookmark it for later. 📖 Watch the course, then read the guide below to reinforce what you learned. What's your favorite Claude feature—Projects, Code, Skills, or Artifacts? 👇 #AI #Claude #Anthropic #PromptEngineering #AIAgents #Productivity
16
180
10
760
75,753
Carlos Marchan retweeted
agents are amazing and have made building software so much easier it’s still so hard to build good software
4
3
12
732
Carlos Marchan retweeted
Every developer thinks they’re “almost done” once they’ve completed about 80% of the work. A lesson I keep learning over and over again: the final 20% usually takes 80% of the time. You ain't done until you are.
79
35
12
485
21,905
Carlos Marchan retweeted
is this a good path to follow?
111
15
3
277
20,410
Now these are breaking … predictions … not that unreal as they would seem at first glance…
16 things that will be normal in 3 years and sound insane today 1. Somebody's entire job is making sure your agents don't do dumb things. 2. Kids grow up assuming any adult who types is old, the way we assume anyone who prints emails is old. 3. You get an itemized bill for what your agents bought last month and it reads like an expense report from a small company. 4. Someone you've never met sells you a business that runs itself, and you never learn what the code does. 5. Your doctor's first opinion comes from a model, and the human's job is deciding whether to trust it. 6. Job posts will say "must be able to manage agents" the same way they used to say "proficient in Excel." 7. The best-paid person at a company will be whoever's best at explaining the business to machines. 8. Your company has more agents than employees, and HR manages both. 9. Companies start hiding how few employees they have, because a lean team reads as fragile to enterprise buyers. 10. Your calendar fills with meetings you didn't schedule, because your agent and their agent worked it out. 11. The CV dies and gets replaced by a body of work an agent can verify in 4 seconds. 12. You interview an agent before you hire it. Give it a fake task, watch how it handles the weird cases, then decide. 13. Losing your job means losing your agents too, and it's kinda scary losing your best agents. 14. "Made by a person" becomes a label on products, and there's a certification body for it. 15. You'll have a folder of agents the way you have a folder of apps 16. Getting a human on the phone becomes a paid tier, and people gladly pay it.
73
Carlos Marchan retweeted
Anthropic just released a free course on loop engineering with Fable 5 00:00 - how Claude Code works under the hood 05:01 - the agentic loop explained 16:21 - the function that 99% of devs ignore 19:01 - why voice beats writing 32:34 - automatic code review with draft PRs 58:39 - Fable 5 for work that's not code This free course replaces any paid Claude Code tutorial. Bookmark it for later.
10
63
1
450
69,856
Carlos Marchan retweeted
We're extending access to Claude Fable 5 on all paid plans through July 12.
5,560
8,937
8,235
80,391
22,933,904
Carlos Marchan retweeted
> … [W]e keep finding things that are mysterious, even unsettling. We find structures that mirror results from human neuroscience. We find evidence of introspection. We find internal states that functionally mirror joy, satisfaction, fear, grief, and unease. I don’t know what that means, but I think it warrants ongoing discernment. > We need more of the world—religious communities, civil society, scholars, governments, and indeed all people of good will … to take this seriously, to look closely, and to push events in a better direction. We need informed critics who will tell the labs when we are failing. We need moral voices that the incentives cannot bend.
Anthropic co-founder Chris Olah was invited to speak at today's presentation of Pope Leo XIV's encyclical "Magnifica humanitas." Read the full text of his remarks: anthropic.com/news/chris-ola…
275
188
52
2,342
375,443
I built "zero2claude", a free course that takes people from zero terminal experience to shipping with Claude Code. The curriculum goes from absolute zero → software basics → Claude Code fundamentals → advanced usage. No shortcuts, no assumptions. 17,000+ students. 7 languages. ~500 active hourly. No marketing. No ads. Pure word of mouth. The entire platform? Built and operated by one person with Claude Code. Lighthouse audit: ✅ Performance : 96 ✅ Accessibility : 100 ✅ Best Practices : 100 ✅ SEO : 100 Production stats: ~6.4M requests/day. 74 req/sec sustained. <0.003% error rate. Claude Code doesn't just write code. It builds production-grade, scalable products. The best way to grow Claude Code adoption isn't to simplify the tool. It's to level up the people. Give fishing rods, not fish. Free. No paywall. My contribution to the community. Link in the comments 👇
61
159
34
1,477
673,525
Carlos Marchan retweeted
The more I replace plans with prototypes, the better the outputs Who'd have thought that low fidelity prototypes were better than walls of spec Oh yeah, the entire industry for 20 years Stop going against decades of knowledge because someone in SF shipped it as a 'mode'
i never make plans i hate looking at markdown i don't wanna read markdown files i just plan by having it make changes to the code then i look at the code to see what sucks then i prompt again
123
109
37
2,038
337,472
Carlos Marchan retweeted
Building a skill that helps you prototype business logic It builds a TUI to help you shortcut through state transitions SO much better than a spec for providing fine-grained feedback Watch the skills repo for /prototype
26
30
2
812
47,812
Agree. My playground artifacts have been so useful as skills explainers. Seeing how @trq212 describes his own experience I cannot help but give it a serious try myself. Will report how things go!
HTML is the new markdown. I've stopped writing markdown files for almost everything and switched to using Claude Code to generate HTML for me. This is why.
23
Pretty smart
/grill-with-docs is MAGICAL It powers up /grill-me by forcing you to define the random jargon your project uses Then every grilling session afterwards is "magically aligned with the thoughts you have" Love a bit of DDD
19
I knew that my DECISIONS.md was pivotal in the quality of the outputs … :)
I'm starting to think that DDD might be the answer to all of my problems - Model not doing what you want? Shared language - Can't navigate a massive codebase? Bounded contexts with global mapping - Don't know why a decision was made? ADR's It's just so freaking elegant
25
Carlos Marchan retweeted
I edited the intro because I realized I buried the lede originally- The 1M context window is a double-edged sword. It allows Claude to do more complex tasks but it can also leads to more context pollution if you don't manage your session well. This is how you do that:
96
116
18
1,600
274,733