boostt.org

Ask New York— It finally answers back Research Preview boostt.org/owning/manhatt… Every answer is an editorial grounded in NYC public records — deeds, permits, filings, assessments. Every claim carries a citation. Nothing is scraped from the web. Nothing is generated from thin air. Honesty and Privacy is preserved by constitutional design. What you get is the public record, made readable. AI surface on Claude, Gemma, with AI Agent MoR marketplace built on Tempo and MPP Machines Payments Protocol for Human and Agent collaboration Try asking: What's the development pipeline around Atlantic Avenue? Which corridors in Brooklyn are seeing permit activity spike? What did comparable buildings sell for near my site last year? Who are the architects filing the most permits in downtown Brooklyn? How has the sales trend in Bed-Stuy shifted over five years? Which licensed plumbers are most active in Queens right now? Who are the electricians pulling permits in my neighborhood? What trades are working on new construction in the Bronx? Which mechanical contractors are filing in Manhattan this year? Stay tuned for the next video: Ask New York NL interface

And a local VC fund 33east.vc

So many startups in this tiny island. So cool people

Startup meetup Cyprus

Dear Anthropic, please stop treating Claude Code users like beta testers Hitting limits after a few prompts on a paid plan is getting ridiculous

こちらの件、プロジェクトとして GitHub 上にリポジトリを作成しました github.com/hiroaki/Tilia ソースをダウンロードすれば手元で動かせますので、あとはよろしくお願いします👀

@realsigridjin I build my own agents, skill and harness. Not sure its a wise decision but it delivers. Although I am just eyeballing the results at this stage

@realsigridjin lean skill fat harness. sending a fat skill to the god model does not make much sense especially if there is a skill orchestrator.

one of the underrated vid recording from hermes agent meetup seoul most agents still start from vague intent. the user gives a loose goal, the model guesses, writes code, patches around missing requirements, and slowly drifts. that can work for demos, but it is fragile for real work. ouroboros takes the opposite approach. before execution, it turns vague intent into a seed. the seed is not just a prompt. it is a contract: what to build, what not to build, what constraints matter, and how success will be judged. that is why ooo auto matters. when someone says, build me a small game for this weekend, the system should not immediately write files. it should clarify the goal, lock the assumptions, create acceptance criteria, and only then hand off to execution. this is the difference between a fat skill and a fat harness. a fat skill adds more prompt, more rules, and more task-specific behavior. a fat harness builds the structure around the model: state, evaluation, recovery, drift control, and execution boundaries. hermes agent fits as the runtime driver. ouroboros owns the workflow. hermes runs the work. the model can generate outputs, but the harness decides what those outputs are allowed to mean. rlm-forge applies the same idea to reasoning. it is not a new model. it is an execution contract for recursive language model work. the system breaks reasoning into child calls, records fresh evidence, and only allows parent synthesis to use that evidence. traceguard enforces the boundary. if a parent claim is not backed by valid evidence from the current run, it rejects. memory can guide behavior, but it cannot become proof. that is the agent os thesis. so the model still matters. but the next layer of leverage is the harness

@realsigridjin old + expert + 18hr/day (Sundays 12 hr) 365/365 = crushing it

young + expert + ai = profit so accurate!!

@gakonst This is a positive sign. Domain expertise is going to thrive

not sure if just me, but things have started to look a lot more the same: * my spotify recommendations aren't as surprising anymore (yay soundcloud) * every ad in sf (okay echo chamber) * every website (stock ai recs) * every roadmap (vertical integration) anyone else? thoughts?

@vkhosla @VinciPhysics New inventions

Most engineers have never had sufficient access to real physics analysis — simulation was too expensive, slow, and specialized. @VinciPhysics vision isn’t just replacing Ansys; it’s bringing continuous physics reasoning to 100x more engineers and 1000x more simulations in a fraction fof the time.

@_nogu66 Looks awesome. I will give it a spin. Thanks

該当プロジェクトはこちらから確認することができます。（バグ修正中） github.com/nogu66/atlas-g…

Google Maps非依存の地球儀アプリを作成した - OSM - Vite - React OSM（OpenStreetMap）は、アメリカの航空会社も使うオープンな地図データ

@mapconcierge Every block has a story. Now anyone can read it. Soon, anyone can write it. Built on OSM

生成AIによるバイブコーディングで、OpenStreetMap の存在に気づかなかった多くの人たちが、Google Maps の代替手段として知らず知らずのうちにOSMを活用する大きなムーブメントになっているのを見ていると、20年近くデータ整備に関わってきた者として感慨深い。#いいぞもっとやれ #OpenStreetMap #古橋研究室 #地球社会共生学部 #AoyamaGSC #一億総伊能化

@chinyich Impressive. Well done

Today, we’re launching illoca Tracing Paper and announcing our $13M Seed round led by Bessemer Venture Partners.

How to apply AI in spatial intelligence? Thats what Ask New York solves ✅Claude does interpretation (the open-ended generalization). The validator does enforcement (the consistency). Together they're a better system than either alone. 📍A pure-rules system can't handle the input variety. 📍A pure-Claude system can't handle the consistency need. This pattern is everywhere in real production AI: Self-driving cars: deep learning for perception (which pixels are a pedestrian?), deterministic logic for control (when to brake). Nobody hands the steering directly to the model Every block has a story. Now anyone can read it. Soon, anyone can write it

Ask New York. It finally answers back Perplexity for spatial intelligence? We are lauching soon and we will find out

@akitec1 cool work. Huge fun of Grasshopper.

Grasshopperを完全なAI Drivenなツールにもっていくためにキャンバスの状態を定義する独自の言語とCLIを作っている...

@madsf88 I'm not sure how your code works under the hood, but if you're open to it — Ask New York on boostt.org could be a fit. A coffee rating map would slot right in. Might benefit New Yorkers. Happy to explore it if you're interested

@madsf88 Really good. congrats

i made strava for coffee. i was on a mission to find the best coffee in nyc and started tweeting my ratings. figured everyone else should be able to do the same. rate cafes, share them, & see where your friends are going

@paulg about garages and boostt.org digital city x.com/paulg/status/2… Over-regulation, including zoning, is a well-documented issue in Europe. The 28th Regime is an attempt to solve fundamental regulatory fragmentation — I'm sure you're aware of it. europarl.europa.eu/thinktank/en/d… I started building boostt.org (28.09.2023) as a layer over cities to allow cross-country service agreements and transactions. How? Service agreements under US law, MoR functionality using Stripe, and Machine Payments Protocol for stablecoin and bitcoin. The digital city has layers The Founders layer Founders are on the map — what they build, what they need, what they offer. The city becomes the incubator. Three marketplaces Workspaces — Housing — Services — founders sell their skills to support their journey Sometime along this two-year journey, the EU-Inc initiative emerged — validating the same thesis from the institutional side. Where I am now? Opening up the city to the world.

@SamCole @paulg thats true

@paulg We started FitXR in a cold, wet garden shed in Woking (just outside of London) All these West Coast garages look luxurious in comparison

It could actually be a significant problem that Europe doesn't have enough garages. This sounds like a joke, but I'm serious. Garages let you work on stuff that doesn't matter yet, which is how big things often start. The outliers of ideas need the outliers of space.