Skip to content
https://abc.microfintool.com/

Mobile & Accessories

Smartphone Reviews, Specs, Comparisons & Electronics Guides

  • Welcome to ABC Tool: Your Ultimate Portal for Smartphones and Accessories
  • PRIVACY POLICY
  • Blog
  • Contact
AddyOsmani.com – Agentic Engineering

AddyOsmani.com – Agentic Engineering

Posted on May 8, 2026October 10, 2026 By Muhammad Safdar No Comments on AddyOsmani.com – Agentic Engineering
Blog

Everything you need to know about addyosmani com agentic engineering.

[ad_1]

A year ago, Andrej Karpathy coined “vibe coding” to describe a gleefully reckless way of programming: you prompt, hand the keyboard to an AI, accept everything it spits out, don’t read the diffs, iterate by pasting error messages back in. It was a great label for a real thing – building quick prototypes or MVPs on pure AI autopilot.

The problem is that “vibe coding” has become a suitcase term. People now use it to describe everything from a weekend hack to a disciplined engineering workflow where AI agents handle implementation under human oversight. These are fundamentally different activities, and conflating them is causing real confusion – and real damage.

What vibe coding actually is

Vibe coding means going with the vibes and not reviewing the code. That’s the defining characteristic. You prompt, you accept, you run it, you see if it works. If it doesn’t, you paste the error back and try again. You keep prompting. The human is a prompt DJ, not an engineer.

This is genuinely useful for:

  • Greenfield MVPs, prototypes and hackathon demos. You need something working by Sunday. Code quality is irrelevant.
  • Personal scripts and one-off tools. You’re the only user. If it breaks, you regenerate it.
  • Learning and exploration. Newcomers can build things they couldn’t otherwise, learning by example from the AI’s output.
  • Creative brainstorming. Deliberately over-generating to see what approaches the AI suggests, then throwing it away and building properly.

If vibe coding gives millions of people the ability to create custom software who otherwise couldn’t, that’s a genuine win. The technique has a legitimate place in the toolbox.

But the failure modes are well-documented at this point. The pattern is always the same: it demos great, then reality arrives. You try to modify it, scale it, or secure it, and you discover nobody understands what the code is actually doing. As one engineer put it, “This isn’t engineering, it’s hoping.”

We need a better term for the professional version

Here’s the thing: a lot of experienced engineers are now getting massive productivity gains from AI – 2x, 5x, sometimes more – while maintaining code quality. But the way they work looks nothing like vibe coding. They’re writing specs before prompting. They’re reviewing every diff. They’re running test suites. They’re treating the AI like a fast but unreliable junior developer who needs constant oversight. I’ve personally liked “AI-assisted engineering” and have talked about how this describes that end of the spectrum where the human remains in the loop.

Simon Willison (whose work I adore) proposed “vibe engineering” for this – it reclaims “vibe” while adding “engineering” to signal discipline. But after watching the community debate this for months, I think the the word “vibe” carries too much baggage. It signals casualness. When you tell a CTO you’re “vibe engineering” their payment system, you can see the concern on their face.

Andrej Karpathy suggested “agentic engineering” this week and I think I like it.

Here’s perhaps why it works:

It describes what’s actually happening. You’re orchestrating AI agents – coding assistants that can execute, test, and refine code – while you act as architect, reviewer, and decision-maker. You might write only a % of the code by hand. The rest comes from agents working under your direction. That’s agentic. And you’re applying engineering discipline throughout. That’s engineering.

It’s professionally legible. “Agentic engineering” sounds like what it is: a serious engineering discipline involving autonomous agents. You can say it to your VP of Engineering without embarrassment. You can put it in a job description. You can build a team practice around it.

It draws a clean line. Vibe coding = YOLO. Agentic engineering = AI does the implementation, human owns the architecture, quality, and correctness. The terminology itself enforces the distinction.

What agentic engineering looks like in practice, perhaps.

The workflow isn’t complicated, but it requires discipline that vibe coding explicitly abandons:

You start with a plan. Before prompting anything, you write a design doc or spec – sometimes with AI assistance. You break the work into well-defined tasks. You decide on the architecture. This is the part vibe coders skip, and it’s exactly where projects go off the rails.

You direct, then review. You give the AI agent a well-scoped task from your plan. It generates code. You review that code with the same rigor you’d apply to a human teammate’s PR. If you can’t explain what a module does, it doesn’t go in.

You test relentlessly. The single biggest differentiator between agentic engineering and vibe coding is testing. With a solid test suite, an AI agent can iterate in a loop until tests pass, giving you high confidence in the result. Without tests, it’ll cheerfully declare “done” on broken code. Tests are how you turn an unreliable agent into a reliable system.

You own the codebase. You maintain documentation. You use version control and CI. You monitor production. The AI accelerates the work, but you’re responsible for the system.

Teams doing this well often report faster development – and those gains come from augmenting a solid process, not abandoning one. The AI handles boilerplate and grunt work. The human focuses on architecture, correctness, edge cases, and long-term maintainability.

The irony is that AI-assisted development actually rewards good engineering practices more than traditional coding does. The better your specs, the better the AI’s output. The more comprehensive your tests, the more confidently you can delegate. The cleaner your architecture, the less the AI hallucinates weird abstractions. As one analysis noted, “AI didn’t cause the problem; skipping the design thinking did.”

The skill gap we’ve discussed

Here’s an uncomfortable truth from the trenches: agentic engineering disproportionately benefits senior engineers. If you have deep fundamentals – you understand system design, security patterns, performance tradeoffs – you can leverage AI as a massive force multiplier. You know what good code looks like, so you can efficiently review and correct AI output.

But if you’re junior and you lean on AI before building those fundamentals, you risk a dangerous skill atrophy. You can produce code without understanding it. You can ship features without learning why certain patterns exist. Several engineering leaders have flagged this as an emerging crisis: a generation of developers who can prompt but can’t debug, who can generate but can’t reason about what they’ve generated.

This isn’t an argument against AI-assisted development. It’s an argument for being honest about what it demands. Agentic engineering isn’t easier than traditional engineering – it’s a different kind of hard. You’re trading typing time for review time, implementation effort for orchestration skill, writing code for reading and evaluating code. The fundamentals matter more, not less.

Where we go from here

The trajectory is clear: AI agents are getting more capable, and the agentic engineering workflow is becoming default for a growing number of professional developers. This is going to accelerate.

We need:

  • Honest terminology. Call it agentic engineering when you mean disciplined, agent-assisted development with human oversight. Call it vibe coding when you mean the fun, reckless, prototyping-only version. Stop using one term for both.
  • Better evaluation frameworks. We need systematic ways to measure whether AI-assisted workflows are actually producing reliable software, not just faster software.
  • Investment in fundamentals. As AI handles more implementation, the premium on architectural thinking, security awareness, and systems design goes up, not down. Training programs need to adapt.

The rise of AI coding doesn’t replace the craft of software engineering – it raises the bar for it. The developers who’ll thrive aren’t the ones who prompt the fastest. They’re the ones who think the clearest about what they’re building and why, then use every tool available – including AI agents – to build it well.

Vibe coding showed us what’s possible when you drop all conventions.

Now it’s time to bring the engineering back. Let’s call that what it is.


I’ve written a new book with O’Reilly, Beyond Vibe Coding, that goes deeper into practical frameworks for AI-Assisted (and agentic) engineering. If you’ve been figuring this out in your own workflow, I’d love to hear what’s working.


[ad_2]
Source link

More on addyosmani com agentic engineering

Stay tuned for updates on addyosmani com agentic engineering as the story develops. Related: Mobile & Accessories.

Frequently Asked Questions

What is the difference between vibe coding and agentic engineering?

Vibe coding means going with the vibes and not reviewing the code: you prompt, accept everything, and iterate by pasting errors back. Agentic engineering means orchestrating AI agents that execute, test, and refine code while the human acts as architect, reviewer, and decision-maker, applying engineering discipline throughout.

Who coined the term vibe coding?

Andrej Karpathy coined vibe coding a year ago to describe a gleefully reckless way of programming where you hand the keyboard to AI, accept everything it spits out, and do not read the diffs.

When is vibe coding actually useful?

It is genuinely useful for greenfield MVPs and prototypes, personal scripts and one-off tools where you are the only user, learning and exploration, and creative brainstorming where you deliberately over-generate and then build properly.

What is the main failure mode of vibe coding?

It demos great, then reality arrives. When you try to modify, scale, or secure the code, you discover nobody understands what it is actually doing. As one engineer put it, this is not engineering, it is hoping.

Related Articles

  • An Ex-Meta L8’s Agentic Engineering Setup
  • The Organizational Playbook for Engineering Transformation at Scale
  • Why is Meta destroying its engineering organization?
Views: 65

Post navigation

❮ Previous Post: Huawei MatePad Pro Max design, key features revealed ahead of launch
Next Post: Samsung’s Galaxy Watches Could Alert Users Before They Faint ❯

You may also like

Trump admin’s coal investments assist plants with repeated violations
Blog
Trump admin’s coal investments assist plants with repeated violations
June 21, 2026
Which Macs are suffering from shortages—and where are things getting worse?
Blog
Which Macs are suffering from shortages—and where are things getting worse?
May 10, 2026
Bear spray is exploding in the trash near Yellowstone National Park
Blog
Bear spray is exploding in the trash near Yellowstone National Park
May 17, 2026
Google Photos could soon give you more tools to make your Memories shine
Blog
Google Photos could soon give you more tools to make your Memories shine
May 30, 2026

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Recent Posts

  • OnePlus 16 Launch Date, Price, Specs: Everything Confirmed for 2026
  • Best 5G Phones 2026 for USA, UK & Europe (Every Budget)
  • Best Foldable Phones 2026: Z Fold 8 vs Pixel 11 Pro Fold vs Razr
  • Google Pixel 10 Review (2026): Still Worth $799?
  • iPhone 16 vs iPhone 17: Should You Upgrade in 2026? (Honest Answer)

Recent Comments

  1. I Hiked Using Robot Legs in the Grand Canyon. I Didn’t Even Need My Cane - Mobile & Accessories on Can Using the Hypershell Exoskeleton on a Bike Replace an E-Bike? I Tested It to Find Out
  2. The media starts receiving the first Trump Mobile T1 Phone units, launch is imminent - Mobile & Accessories on The Competitive Moat That AI Can’t Replicate
  3. Samsung’s giant The Frame Pro TV is $1,500 off until midnight - Mobile & Accessories on All Your Wearables Have One Glaring Weakness. What Can You Do About It?
  4. iPhone 18: Everything We Know About Apple's Most Ambitious Lineup Yet - Mobile & Accessories on This magnetic Qi 2.2 power bank is perfect for the iPhone 17 and it’s on sale right now
  5. Do Camera Sensor Sizes Matter? - Mobile & Accessories on Xiaomi 18 Pro set for a screen upgrade that Samsung fans will be very familiar with

Archives

  • October 2026
  • June 2026
  • May 2026
  • April 2026

Categories

  • Blog
  • Mobiles

Copyright © 2026 Mobile & Accessories.

Theme: Oceanly News by ScriptsTown