frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Ask HN: Who's really winning with AI-Coding?

2•unclebucknasty•7m ago
I've noticed the same pattern across three different Claude-coded projects now. Initial velocity is strong, but starts to fade until returns are negative.

It seems a combination of changes in model behavior and increasing code base size—emphasis on the latter. The larger the code base and documented history/progress (e.g. through updates to completed work and instructions, as in .md files), the more drift starts to set in.

At a certain point, you get collapse. This is where it can't maintain consistency within a single unit of work. It can manifest, for example, as adding a similar element to a view, but styling it in a completely different manner than existing elements. It literally feels random.

It can also manifest as incoherent or poorly implemented logic and behavior. I've followed different advice, restructured projects, entered plan mode, added deterministic checks, styles, rules, etc. But, it feels like whack-a-mole and it's unclear that it actually adheres to these rules in the first place.

Then, there's the question of it moving the problem. In response to my complaints of inconsistent UI generation, it recommended a style check to run after its UI changes. Of course, that begs the question of why it doesn't just apply those rules in the initial generation.

But, I am at a place now where it spins up AI agents to do a full UI/UX review—Playwright and all. Yet, it's still drifting and not catching it. This is consistent with its unit tests—on my latest project it has now generated a suite of hundreds of somewhat time consuming tests that it constantly runs pre-commit, and not a single test has ever failed. Meanwhile, there have been a number of bugs. When questioned, it said it had been testing the "wrong layer" and had essentially generated self-confirming tests.

My question is who is really winning here? When querying Google, it points to so many claims of success and approaches that supposedly contribute to that success. But, when you break down the underlying source for those claims, it appears to be largely Reddit.

Yet, pressing it further reveals that virtually every actual study shows either little to no productivity gains or actual regression—particularly where experienced engineers are concerned.

So, who here has developed a quality production application of any complexity using AI coding? If anyone, which harness/model, and what techniques did you use? Also, can you point us to your repo?

Handing over Humanity

https://www.writingruxandrabio.com/p/handing-over-humanity
1•gmays•31s ago•0 comments

AI Lives Don't Matter

https://www.ailivesdontmatter.com/
1•logn•2m ago•0 comments

Tips for beginner solo backpackers (2023)

https://beeswildlife.com/2023/12/16/5-tips-for-beginner-solo-backpackers/
1•mooreds•2m ago•0 comments

DOJ Willing to Clear More Merger Deals Without Asset Sales

https://www.bloomberg.com/news/articles/2026-11-16/trump-doj-willing-to-clear-more-merger-deals-w...
1•petethomas•3m ago•0 comments

Ask HN: Who's really winning with AI-Coding?

2•unclebucknasty•7m ago•1 comments

Demolition Van – I've took the most difficult VC mission to extreme

https://demolitionvan.com/
1•marrrcin•8m ago•1 comments

Coding Machines

https://www.teamten.com/lawrence/writings/coding-machines/
1•fedpost•8m ago•0 comments

GTA set in the full map of San Francisco

https://gta.mehranjalali.com
1•MehranJ•10m ago•0 comments

Show HN: Doku, a browser-based Sudoku app I built

https://doku-psi.vercel.app/
1•luvielmo•12m ago•0 comments

Dumping an AES key from a MyQ garage opener to stop rebooting when offline

https://kusha.me/blog/stop_myq_homebridge_819lmb_rebooting/
2•kusha•12m ago•1 comments

Show HN: Listodo – A to-do list with an AI agent for every task

https://github.com/debarshibasak/listodo
1•debarshri•14m ago•0 comments

I'm a full-time NHS doctor. I built my clinic's platform with Claude Code

https://lately.health/
2•shadasinger•17m ago•0 comments

Tesla Cybercab has crashed 8 times, and the data shows a speed problem

https://electrek.co/2026/11/16/tesla-cybercab-crashes-speed-problem/
3•cdrnsf•18m ago•1 comments

Uber Adds Zoox's Robotaxis to Its App. Everything to Know

https://www.cnet.com/roadshow/automobiles/uber-adds-zoox-robotaxis-to-app/
1•sshah03•19m ago•0 comments

Zero-Upload Data Cleaner: Clean Messy CSV, Parquet, and JSON

https://lopespm.com/p/cleaner/
1•lopespm•23m ago•0 comments

Flawless UI-UX Skill 3.0.0

https://github.com/smokeeaasd/flawless-ui-ux
4•nkwwzn•24m ago•0 comments

EWD 1036 – On the cruelty of teaching computing science (1988)

https://www.cs.utexas.edu/~EWD/transcriptions/EWD10xx/EWD1036.html
1•ky3•24m ago•0 comments

New privacy rules for Google's Gemini Apps (updated November 16, 2026)

https://support.google.com/gemini/answer/13594961?hl=en
1•tedlistens•31m ago•0 comments

Show HN: Sutra-OS – A self-hosted ontology platform

https://github.com/0nelight/sutra-os
9•0nelight•34m ago•0 comments

California sues OpenAI over hacking risks

https://www.npr.org/2026/11/16/nx-s1-6042920/california-sues-openai-hugging-face-hacking-incident
5•CSSer•37m ago•0 comments

The Robotaxi Blame Game: Who Goes to Jail When the Autonomous Car Crashes?

https://2tinteractive.com/watch/episodes/5641/
2•LebToki•37m ago•0 comments

Misalignment signals in Kimi K3: What probes flag and where they peak

https://commandline.microsoft.com/misalignment-signals-in-kimi-k3-what-probes-flag-and-where-they...
1•MarcoDewey•40m ago•0 comments

The FAA's ground stop may have a simple culprit. Its causes may not

https://www.fastcompany.com/91643272/the-faa-was-warned-about-its-telecom-upgrade
2•tedlistens•40m ago•0 comments

Noeri: A neolab to push the frontier of plural intelligence

https://noeri.ai/
1•polymorph1sm•41m ago•0 comments

Performance Boost for Intel CPUs: Impact of FRED Measured

https://www.hwcooling.net/en/performance-boost-for-intel-cpus-impact-of-fred-measured/
1•teleforce•44m ago•0 comments

Making a Staff Sling

https://www.youtube.com/watch?v=zSHD8RG_mSo
2•lifeisstillgood•46m ago•0 comments

Rust: Enabling the parallel front end on nightly

https://blog.rust-lang.org/2026/11/16/enabling-parallel-frontend-on-nightly/
2•matt_d•46m ago•0 comments

What Comes After HBM

https://www.mackenziemorehead.com/what-comes-after-hbm/
1•gmays•47m ago•0 comments

The Name Is the Spec

https://fromautomation.co/p/the-name-is-the-spec
1•ghels1444•47m ago•0 comments

Microsoft Bug Undetected Since July Triggered Worldwide Outage

https://www.bloomberg.com/news/articles/2026-11-16/microsoft-bug-undetected-since-july-triggered-...
4•rndsignals•50m ago•0 comments