frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: Dumped Wix, my AEC consultancy's storefront is now an AI Edge

https://axoworks.com/
1•axotopia•1h ago
I run a building design consultancy for homeowners and architects, not a SaaS firm. Honestly, I'm not going to claim we were trying to build some fantastic ‘anti-fragile alternative’ for the future… I just got tired of paying Wix $40 a month for a brochure no one read. The portfolio was static, the inquiries were generic, and every time a potential client asked about setback variances, I'd lose hours explaining the same thing. So last December, I told my wife I was killing the website. She thought I’d lost it. Maybe I had.

The stack? It’s messy. We started on Netlify because it was cheap. Hit that stupid 10-second serverless timeout after a week when I started customizing the bot. The bot would just… freeze. Had to implement a clunky polling loop that felt like a band-aid. Eventually split the agent into 3 pieces just to keep it responsive:

The Brain (Edge) figures out what you want, shoots a JSON signal to the browser. The Hands (Browser) runs the actual tool—like pulling a code reference—via a separate serverless function. The Voice (Edge) feeds the result back to talk.

It works, but it’s not elegant. It’s duct tape. Last time I coded was over 30 yrs ago. AI helped a lot, but it feels like 3 steps forward, 2 steps back when it starts obsessing about one area and wrecks what was done 5 refactors ago.

DeepSeek-R1 is the reasoning engine. We tried Gemini and a bunch of others initially. They were either too chatty or horribly bone dry. DeepSeek stays on script. Interestingly, I discovered a couple of weeks ago that a traditional architect—probably not liking me putting a dent in his ‘protected’ professional business—tried to pick a fight with the bot. Wanted to prove it couldn’t handle “professional nuance.” DeepSeek shut him down without breaking a sweat. Log is public if you want to see it.

Fallback is MiniMax M2.5. Triggers on rate limits. Sometimes it feels a bit slower, but it hasn’t failed yet.

Voice input via Web Speech API. That was a mistake. Transcription hallucinations are a nightmare. “Setback” becomes “set back” becomes “backset.” We added a dual-path validation layer. Speech-to-text gets audited before the reasoning models even see it. Still not perfect.

Liability is the real killer. In our world, if the AI hallucinates a building code clause, we’re done. Insurance won’t touch us. We’ve gone through four architectural versions. Started with 30k lines of spaghetti. Now it’s decoupled, but still fragile. We publish the crash logs publicly because hiding them feels dishonest. Also, it keeps us honest.

The hardest part wasn’t the infrastructure, the intent engineering was. Making an LLM sound like a seasoned principal when talking to a first-time homeowner, then pivot instantly to defend our biz model against an angry architect who wants to punch holes… that took 2 whole months.

Speed was killing us. So I came up with this hack—I call it ‘Eager RAG’ because it’s basically guessing what you’ll ask to see if we have something similar already in the database of controlled responses to cut response time. It burns through tokens like crazy, but man, it makes the thing feel much more instant. We also ripped out the persistent databases. Turns out 19 out of 20 visitors never come back, so why bother? Feels wasteful until you realize you’re not building for the 5% that return.

It’s not flawless. No server-side job queue yet. If a client drops mid-query, the result vanishes. But it lets me operate without junior staff. That’s the point.

The terminal is live at https://axoworks.com. Mic optional. No signup. Try to break it. Ask a complex zoning question. See if it over-promises.

Logs:

Architect vs. DeepSeek: https://logs.axoworks.com/chat-architect-vs-concierge-v147.h...

System Audit: https://logs.axoworks.com/audit-2026-02-19-v148.html

I’ll be in the comments. Thanks. Kee

UMD Scientists Create 'Smart Underwear' to Measure Human Flatulence

https://cbmg.umd.edu/news-events/news/brantley-hall-umd-scientists-create-smart-underwear-measure...
1•ohjeez•42s ago•0 comments

Weak-Form Evolutionary Kolmogorov-Arnold Networks for Solving PDEs

https://arxiv.org/abs/2602.18515
1•PaulHoule•1m ago•0 comments

Compile TypeScript directly to JVM bytecode and run it at native Java speed

https://blog.caoccao.com/compile-typescript-to-jvm-bytecode-with-swc4j-45dd05561028
2•sjtucaocao•1m ago•1 comments

A preview of Coalton 0.2, a statically-typed Lisp

https://coalton-lang.github.io/20260312-coalton0p2/
2•fanf2•4m ago•0 comments

Show HN: GDSL – 800 line kernel: Lisp subset in 500, C subset in 1300

https://firthemouse.github.io/
1•FirTheMouse•5m ago•0 comments

Built a 1.3M-line agent-native OS in Rust while homeless. What now?

2•jamieoglindsey•6m ago•0 comments

Bunnie's Piggyback Hack

https://pluralistic.net/2026/01/09/quantity-break/
1•cat-whisperer•6m ago•0 comments

Turing Drawings (2012)

https://maximecb.github.io/Turing-Drawings/
1•tliltocatl•6m ago•1 comments

Will the Indus Valley script ever be deciphered?

https://www.livescience.com/archaeology/will-the-indus-valley-script-ever-be-deciphered
1•Brajeshwar•7m ago•0 comments

Show HN: Detect Drift in TypeScript Codebases

https://github.com/LogicStamp/logicstamp-context
1•AmiteK•7m ago•0 comments

Arduino Ventuno Q First Look: Benchmarks, Specs and Mainline Linux

https://sbcwiki.com/news/articles/arduino-ventuno-q-first-look-ew26/
1•HeyMeco•8m ago•0 comments

Neuroglancer: WebGL-based viewer for volumetric data

https://github.com/google/neuroglancer
1•cl3misch•8m ago•1 comments

'Pokémon Go' players have been unknowingly training delivery robots

https://www.popsci.com/technology/pokemon-go-delivery-robots-crowdsourcing/
2•jonbaer•9m ago•0 comments

AI agent hacks McKinsey: five situations when you should not deploy agents

https://nanonets.com/blog/ai-agent-hacks-mckinsey/
2•Terretta•10m ago•0 comments

Tokenization Is Killing Our Multilingual LLM Dream

https://huggingface.co/blog/omarkamali/tokenization
1•omneity•11m ago•0 comments

I Just Wanted a Storage Bin. I Built a Search Engine Instead

https://binsizes.com/blog/story
2•ryanscio•12m ago•0 comments

Most of the AI software YC companies are coming out with doesn't work

1•somethingtoday•13m ago•0 comments

Minimally invasive optical clearing media: live cell imaging ex vivo and in vivo

https://www.nature.com/articles/s41592-026-03023-y
1•bookofjoe•13m ago•0 comments

Overthinking

https://kerkour.com/overthinking
1•randomint64•14m ago•0 comments

Digg's open beta shuts down after just two months, blaming AI bot spam

https://www.theverge.com/tech/894803/digg-beta-shutdown-layoffs-ai
1•evah•14m ago•0 comments

$3 AliExpress Heart Rate and Oxygen Sensor Module

https://www.youtube.com/watch?v=0OV5aCOnXBA
1•iamflimflam1•16m ago•0 comments

Trump call for Japan to send warships to Strait of Hormuz puts Takaichi in bind

https://www.japantimes.co.jp/news/2026/03/15/japan/politics/trump-hormuz-strait-japan-takaichi/
3•mikhael•18m ago•1 comments

Pretraining Language Models via Neural Cellular Automata

https://hanseungwook.github.io/blog/nca-pre-pre-training/
2•shmublu•22m ago•0 comments

UK to Allow "Plug in" Solar

https://www.gov.uk/government/news/government-to-go-further-and-faster-in-becoming-energy-secure
2•edent•23m ago•0 comments

Don't Call It a Moat

https://twitter.com/yrechtman/status/2031770436309029227
1•gmays•23m ago•0 comments

Block Number Formats are (Still!) Direction Preservers

https://constantinides.net/2026/03/15/block-number-formats-are-still-direction-preservers/
1•matt_d•24m ago•0 comments

Show HN: Nom – A public feed of your GitHub activity, auto-summarized

https://nomit.dev/
1•Lws803•24m ago•1 comments

Learning Athletic Humanoid Tennis Skills from Imperfect Human Motion Data

https://zzk273.github.io/LATENT/
1•danielmorozoff•25m ago•0 comments

Couplit: Daily coupled word-shuffle puzzle

https://couplit.io/
1•yig•25m ago•0 comments

Ask HN: How do you handle personal finance without giving data to third parties?

1•francesco_gab•26m ago•0 comments