Devesh Mankar | @[email protected]

Turning distributed systems experience into AI architecture.
@Thoughtworks Tech Lead → Tech Principal | Ex @BMCSoftware
Writing about what I learn. No fluff.

[Devesh Mankar | @[email protected] @Devz15]

Most people use AI coding agents like chatbots.
That’s why they get mediocre results.

The real win is in planning, memory, delegation, compaction, and verification.

Works for Claude Code, Copilot, Amazon Q, Cursor—basically any serious AI coding stack.

https://t.co/c037dh9MKf

#LatePost #AI #CodingAssistant #Productivity #DevTools

[Devesh Mankar | @[email protected] @Devz15]

@karpathy nails the shift here, but one nuance:

We’re misidentifying what part of software engineering is being automated.

LLMs are great at generating code. But systems don’t fail at generation - they fail at assumptions, edge cases, and correctness.

The “no IDE needed” narrative misses this completely.
IDEs are where engineers review, validate, and reason about production systems.

Anyone saying otherwise is simply farming engagement and after views, impressions and likes.

#LLMs #AIEngineering #SoftwareEngineering

[Andrej Karpathy @karpathy]

A few random notes from claude coding quite a bit last few weeks.

Coding workflow. Given the latest lift in LLM coding capability, like many others I rapidly went from about 80% manual+autocomplete coding and 20% agents in November to 80% agent coding and 20% edits+touchups in December. i.e. I really am mostly programming in English now, a bit sheepishly telling the LLM what code to write... in words. It hurts the ego a bit but the power to operate over software in large "code actions" is just too net useful, especially once you adapt to it, configure it, learn to use it, and wrap your head around what it can and cannot do. This is easily the biggest change to my basic coding workflow in ~2 decades of programming and it happened over the course of a few weeks. I'd expect something similar to be happening to well into double digit percent of engineers out there, while the awareness of it in the general population feels well into low single digit percent.

IDEs/agent swarms/fallability. Both the "no need for IDE anymore" hype and the "agent swarm" hype is imo too much for right now. The models definitely still make mistakes and if you have any code you actually care about I would watch them like a hawk, in a nice large IDE on the side. The mistakes have changed a lot - they are not simple syntax errors anymore, they are subtle conceptual errors that a slightly sloppy, hasty junior dev might do. The most common category is that the models make wrong assumptions on your behalf and just run along with them without checking. They also don't manage their confusion, they don't seek clarifications, they don't surface inconsistencies, they don't present tradeoffs, they don't push back when they should, and they are still a little too sycophantic. Things get better in plan mode, but there is some need for a lightweight inline plan mode. They also really like to overcomplicate code and APIs, they bloat abstractions, they don't clean up dead code after themselves, etc. They will implement an inefficient, bloated, brittle construction over 1000 lines of code and it's up to you to be like "umm couldn't you just do this instead?" and they will be like "of course!" and immediately cut it down to 100 lines. They still sometimes change/remove comments and code they don't like or don't sufficiently understand as side effects, even if it is orthogonal to the task at hand. All of this happens despite a few simple attempts to fix it via instructions in CLAUDE . md. Despite all these issues, it is still a net huge improvement and it's very difficult to imagine going back to manual coding. TLDR everyone has their developing flow, my current is a small few CC sessions on the left in ghostty windows/tabs and an IDE on the right for viewing the code + manual edits.

Tenacity. It's so interesting to watch an agent relentlessly work at something. They never get tired, they never get demoralized, they just keep going and trying things where a person would have given up long ago to fight another day. It's a "feel the AGI" moment to watch it struggle with something for a long time just to come out victorious 30 minutes later. You realize that stamina is a core bottleneck to work and that with LLMs in hand it has been dramatically increased.

Speedups. It's not clear how to measure the "speedup" of LLM assistance. Certainly I feel net way faster at what I was going to do, but the main effect is that I do a lot more than I was going to do because 1) I can code up all kinds of things that just wouldn't have been worth coding before and 2) I can approach code that I couldn't work on before because of knowledge/skill issue. So certainly it's speedup, but it's possibly a lot more an expansion.

Leverage. LLMs are exceptionally good at looping until they meet specific goals and this is where most of the "feel the AGI" magic is to be found. Don't tell it what to do, give it success criteria and watch it go. Get it to write tests first and then pass them. Put it in the loop with a browser MCP. Write the naive algorithm that is very likely correct first, then ask it to optimize it while preserving correctness. Change your approach from imperative to declarative to get the agents looping longer and gain leverage.

Fun. I didn't anticipate that with agents programming feels *more* fun because a lot of the fill in the blanks drudgery is removed and what remains is the creative part. I also feel less blocked/stuck (which is not fun) and I experience a lot more courage because there's almost always a way to work hand in hand with it to make some positive progress. I have seen the opposite sentiment from other people too; LLM coding will split up engineers based on those who primarily liked coding and those who primarily liked building.

Atrophy. I've already noticed that I am slowly starting to atrophy my ability to write code manually. Generation (writing code) and discrimination (reading code) are different capabilities in the brain. Largely due to all the little mostly syntactic details involved in programming, you can review code just fine even if you struggle to write it.

Slopacolypse. I am bracing for 2026 as the year of the slopacolypse across all of github, substack, arxiv, X/instagram, and generally all digital media. We're also going to see a lot more AI hype productivity theater (is that even possible?), on the side of actual, real improvements.

Questions. A few of the questions on my mind:
- What happens to the "10X engineer" - the ratio of productivity between the mean and the max engineer? It's quite possible that this grows *a lot*.
- Armed with LLMs, do generalists increasingly outperform specialists? LLMs are a lot better at fill in the blanks (the micro) than grand strategy (the macro).
- What does LLM coding feel like in the future? Is it like playing StarCraft? Playing Factorio? Playing music?
- How much of society is bottlenecked by digital knowledge work?

TLDR Where does this leave us? LLM agent capabilities (Claude & Codex especially) have crossed some kind of threshold of coherence around December 2025 and caused a phase shift in software engineering and closely related. The intelligence part suddenly feels quite a bit ahead of all the rest of it - integrations (tools, knowledge), the necessity for new organizational workflows, processes, diffusion more generally. 2026 is going to be a high energy year as the industry metabolizes the new capability.

[Devesh Mankar | @[email protected] @Devz15]

📢 UPDATE:

First, a huge thank you to everyone who amplified this issue. 🙏❤️ The overwhelming response from the community helped bring attention to a genuine transparency "gap" in the repair and replacement process.

Google's higher support team has now stepped in and offered a brand-new replacement device, which I sincerely appreciate. 👏

However,
⚠️ Customers should not have to escalate, go viral, or seek exceptions just to get basic transparency before making a payment decision.

Your support to original post shows that this is not an isolated concern. Many customers clearly want greater transparency around:
📋 Replacement device details
🔋 Battery health standards
💧 Water resistance/IP integrity
✨ Cosmetic condition standards
🔄 Customer options before payment

Providing this information upfront would help customers make informed decisions and strengthen long-term confidence in the Pixel ecosystem and repair process.
🤝 Transparency builds trust.
📈 Trust builds loyalty.
📱 Loyalty builds brands.

I hope @Google uses this feedback to address the process "gap" and improve the experience for all customers - not just those who escalate.
@GoogleIndia @madebygoogle
#GooglePixel #PixelIndia #ConsumerRights #RightToRepaird

[Devesh Mankar | @[email protected] @Devz15]

🚨 WARNING for all Pixel owners in India. 🚨

My Pixel device was sent to official Google repair. After inspection, Google declared it “not repairable” and gave me only 2 choices:

1️⃣ Pay ₹22,620 for a replacement device
2️⃣ Get the phone returned unrepaired

Now comes the shocking part.

Before taking ₹22k+, Google India REFUSES to disclose:

❌ Which exact device they will send
❌ Whether it’s NEW or REFURBISHED
❌ Battery health details
❌ Water Resistance Rating
❌ Replacement quality standards
❌ Google cannot guarantee battery health percentage
❌ Google cannot guarantee IP/water resistance rating

And in writing, Google Support confirmed:
⚠️ Once payment is made, there is NO refund and NO return — even if you are unhappy with the replacement device quality sent later.

Most shocking part?

Google could not even confirm whether they have any minimum measurable quality standards/policies for refurbished replacement devices:

Battery health %
Water resistance integrity like IP rating

So basically:

Pay first.
Find out later what device you got.
No refund. No choice.
Without transparency, measurable quality benchmarks or refund rights.

This is not premium after-sales support.
This is “trust us after payment.”

Attaching screenshots of Google Support’s own email response.

Indian Pixel buyers deserve transparency before payment, not blind acceptance.

@GoogleIndia @madebygoogle
#GooglePixel #PixelIndia #Pixel6 #GoogleIndia #ConsumerRights #RightToRepair #MadeByGoogle #TechTwitter #Android #Smartphones

[Devesh Mankar | @[email protected] @Devz15]

@madebygoogle I have DMed You : Google Repair Request ID : RMA34511142

[Devesh Mankar | @[email protected] @Devz15]

One developer, one leak, one night.

@realsigridjin rewrote Claude Code into Claw Code in just hours—and the repo exploded.

The real story? It wasn’t about faster coding.
It was about orchestration.

https://t.co/JtubUX5lm5

#LatePost #AI #DeveloperTools #OpenSource #GitHub

[Devesh Mankar | @[email protected] @Devz15]

#BlogPost #LatePost #AI #ClaudeCode #OpenSource #AIAgents
Anthropic accidentally exposed the blueprint behind Claude Code. 512,000 lines. 1,900 files. Hidden features. Undercover Mode.

Then the open-source world responded overnight. This wasn’t just a leak. It was a preview of where AI agents are headed.

Read more at https://t.co/d9yMLk68wC

[Tech Bharat (Nitin Agarwal) @techbharat]

I dare @google to formulate such policies for the US customers! Why this cuck behaviour @GoogleIndia @sundarpichai?

[Devesh Mankar | @[email protected] @Devz15]

@ThePuneMirror @grok kindly fact check this

[Devesh Mankar | @[email protected] @Devz15]

I just published Switching Your Pixel Phone to GrapheneOS: The Definitive Guide to What You Gain, What You Lose, and… https://t.co/n4V5k6iJva

[Devesh Mankar | @[email protected] @Devz15]

@jsensarma Due to a youtube video which recommended your handle.. I'm one of those who followed by seeing that video

[Devesh Mankar | @[email protected] @Devz15]

Do you remember when you joined X? I do! #MyXAnniversary

[Devesh Mankar | @[email protected] @Devz15]

@Ghanshamkakde1 Wish him a speedy recovery!!

[Devesh Mankar | @[email protected] @Devz15]

Recency Bias: We tend to favor the ideas, solutions, and influences that occurred to us most recently, regardless of whether they are the best ones! #Insightful #idea #quote

[Devesh Mankar | @[email protected] @Devz15]

Reading the fascinating and relatable book "Building Your Second Brain" by @fortelabs. I'll be sharing key excerpts here as I read along, so stay tuned for insights on how to organize your digital life, learn more effectively, and unlock your creative potential! #SecondBrain

[Devesh Mankar | @[email protected] @Devz15]

Despite all the mind-expanding ideas we have access to, the quality of our attention is only getting worse. We are paralysed by the conflict between our responsibilities and our most heartfelt passions!

[Devesh Mankar | @[email protected] @Devz15]

Information overload has become information exhaustion. Instantaneous access to the world's knowledge through the internet was supposed to educate and inform us but instead it has created a society-wide poverty of attention! #quote #insightful

[Devesh Mankar | @[email protected] @Devz15]

Information overload has become information exhaustion. Instantaneous access to the world's knowledge through the internet was supposed to educate and inform us but instead it has created a society-wide poverty of attention! #quote #insightful

[Devesh Mankar | @[email protected] @Devz15]

Reading the fascinating and relatable book "Building Your Second Brain" by @fortelabs. I'll be sharing key excerpts here as I read along, so stay tuned for insights on how to organize your digital life, learn more effectively, and unlock your creative potential! #SecondBrain

[Devesh Mankar | @[email protected] @Devz15]

Reading the fascinating and relatable book "Building Your Second Brain" by @fortelabs. I'll be sharing key excerpts here as I read along, so stay tuned for insights on how to organize your digital life, learn more effectively, and unlock your creative potential! #SecondBrain

[Tesla Club India® @TeslaClubIN]

"You're Being Lied to About Electric Cars

Science has repeatedly shown EVs are better for humans, despite the meme you just retweeted."

https://t.co/Poo5j3M1Ds

[Devesh Mankar | @[email protected] @Devz15]

This is an easy migration guide to Mastodon https://t.co/cXKqRm3IOM

[Devesh Mankar | @[email protected] @Devz15]

@debirdify searched 431 Twitter accounts that I follow and found 26 accounts with 26 Fediverse IDs, spread over 11 instances! Come on Tweeple, is y high time to join Mastodon! My handle is https://t.co/NBoEZlgAnG , updated on profile too.