This is all new, so I may be a bit hyperbolic, but the reason OpenAI introducing ads bothers me is the implicit (or even explicit) bias that can be smuggled into a chat in ways that simply aren't possible when you're just clicking through to an external source. There are all kinds of implications to Google no longer being that source of truth, even by default. Maybe this has quietly been the case for a long time, but this feels like the final move — pushing their ad bias (i.e., whoever paid the most) into a conversational system, where dark patterns are far easier to implement and much harder to detect.

One answer to this might be domain-specific agents — narrower, accountable, ideally something you (or your community) actually run. But even then it all falls back on trust: you being a good-faith actor, and others trusting that you are one. Which is to say, we're back to the same problem, just at a smaller scale.

Mine is: https://arionhardison.com and the main things I wanted to know about others personal agents is:

1. What is the main purpose of it?

2. Do you and if you do how do you use and/or implement skills?

3. Does your agent utilize HITL and if so what triggers it?

In my case:

1: The main purpose is to replace my email and/or provide a public phone number w/ smart SPAM filtering. 2. It implements the "skills" that I have listed on my resume. 3. Each conversation weights an index and if the weight gets high enough then it will contact me w/ a context summary so I can take over the conversation if needed.

[1] - https://news.ycombinator.com/item?id=19197533

Comments URL: https://news.ycombinator.com/item?id=47953191

Points: 3

# Comments: 0

This is a really good point though, I think I should remove that.

2. I agree, I think I was a bit too worried because I do not know how to navigate this space "Gov Tech" very well.

Thankyou very much for your response; developing a product in a silo can cause tunnel vision which leads to blowing things out of proportion, your comment has really helped me to put things into perspective.

My biggest concern by far is that they seem to have put codify.inc on a registry so ISP's are blocking or showing the red "this is a scam - go back to safety" page. I really liked and invested a lot into that "branding".

I'm Arion. I'm building Project20x — an AI-native governance platform (policy authoring → codification → delivery as digital public goods). It's the substrate that turns policy into running services. I'm building it across all 50 states and 40+ countries concurrently, because government actually runs on interagency dependencies, not silos — VA hands off to HUD, HHS coordinates with every state Medicaid office, etc.

The subdomain pattern at the time was {agency}.{state}.{country}.codify.inc — so the Hawaii subprojects lived at dlir.hi.usa.codify.inc, health.hi.usa.codify.inc, etc. Real staging environments. Not impersonations. No credential capture, no solicitation of money, no fake state seal.

In late 2025 the Hawaii SOC published an alert flagging those subdomains as phishing impersonating state agencies — and pushed it out as a press release. KITV ran the segment in the URL above. Several other Hawaii outlets ran their own write-ups off the same release. So "scam" is now indexed across multiple sites, not one. The same effort that went into a coordinated press push could have gone into one email to a contact page — but I hadn't published one, and they didn't ask.

Here's what I think is fair, and what I think isn't:

Fair: A citizen unfamiliar with Codify could be thrown by a .inc URL that contains an agency abbreviation as a subdomain label (dlir.hi.usa.codify.inc) — even though the apex is codify.inc not .gov, and every page header read "Codify Inc official portal for [agency name]." The on-page identification was there; the URL itself was the surprise. That's a comms-and-onboarding failure on my part, and the fix is to stop putting agency abbreviations into deep subdomain paths. The new pattern is per-city apex (codify.la, codify.nyc, codify.boston, codify.miami, codify.vegas) — clearly a Codify property at first glance, no nested abbreviations to misread. I've also published a security contact (a@project20x.com) and a public registry listing live vs. staging vs. claimable subprojects. The SOC didn't reach out before the alert because I hadn't published a contact. That's fixed.

Not fair: "Scam" is a factual claim and it's wrong. Every page header on the flagged subdomains read "Official Codify Inc portal" or "Official Project20x portal for [agency name]" — including the screenshots used in the "scam" example. The site never claimed to be the agency, never collected information on the agency's behalf, and never solicited money. This is a civic-tech project in the same spirit as DOGE / USDS / 18F — same DOGE-shaped goal, achieved by compilation rather than chainsaw. Building in public on the open internet has a cost I underestimated, but mislabeling civic-tech as fraud has a cost too.

Why I'm asking now: I'm launching the city-tier version — "DOGE for cities" — for Miami, Boston, NYC, LA, and Las Vegas, on per-city apex domains (codify.miami, codify.la, codify.nyc, codify.boston, codify.vegas). No more nested codify.inc subdomains. Playbook this time: clear "Codify Inc portal" header on every page, published security contact, .gov counterpart links, and CIO/CISO outreach before launch. Rather get it right than clean up again.

So — HN, what would you actually do?

Project: https://project20x.com/about Contact: a@project20x.com

Comments URL: https://news.ycombinator.com/item?id=47922126

Points: 1

# Comments: 6

Comments URL: https://news.ycombinator.com/item?id=47889989

Points: 3

# Comments: 0

The idea: describe any problem in plain language (voice or text), and AI codifies it into a structured program with the right people, steps, timeline, and agents to get it done. It's a 5-step wizard: Define Problem → Codify Solution → Setup Program → Execute Program → Verify Outcome.

It runs across 50+ domains — codify.healthcare (EMR backend), codify.education (LMS backend), codify.finance, codify.careers (HRM backend), codify.law, plus 13 city domains (codify.nyc, codify.miami, codify.london, codify.tokyo, etc.). Each domain tailors the AI assessment and program output to that sector.

The platform is Project20x — think of it as the infrastructure layer. If Codify is the verb ("codify your healthcare problem into a care program"), Project20x is the operating system that runs it all: multi-tenant governance, AI agent orchestration, and domain-specific sys-cores for healthcare, education, city services, etc.

Every US federal agency and state-level department has a subdomain — ed.usa.project20x.com (Dept of Education), doj.usa.project20x.com, hhs.usa.project20x.com, etc. — with AI agents representing each agency's mandate. Same structure at the state level.

The political side: Project20x hosts policy management for both parties — dnc.project20x.com and rnc.project20x.com — where legislative intent gets codified into executable governance through a 10-step policy lifecycle. Right now I'm building out the multi-agent environment so agency agents can negotiate with each other, make deals, and send policy proposals up to the HITL (human-in-the-loop) politician for approval. Each elected official has a profile (e.g. https://project20x.com/u/donald-trump) where constituents can engage and where policy proposals land for review.

The name is a nod to structured policy frameworks, but the goal is nonpartisan infrastructure: democratically governed essential services delivered as AI-native social programs.

Stack: Nuxt 2/Vue 2 frontend, Laravel 10 API, Python/LangGraph agent orchestration, Flutter mobile app. Currently live across all domains.

https://project20x.com | https://codify.healthcare | https://codify.education | https://dnc.project20x.com | https://rnc.project20x.com etc...

1. The presidents response to bombing of school girls was basically "stop hitting yourself"

2. Fox news host Dept. of Defense head and the "Dept. of War" name "change"

3. Building a grand ballroom while taking benefits away from hungry kids

4. Elon musk on stage with the chainsaw bragging about acts that save no money but did harm the poorest people on earth.

5. The fact that our media does not really care about any of this unless they get a ratings bump from it

Obviously we all could go on and on.. but the biggest loss IMO is objective truth. There are and will always be things that are true and I feel that we are losing a hold of that so that bad actors can just say to us: "no thats not what your seeing".

Its like in the movie, if they had looked at the plant growing and said: "Thats FAKE NEWS" then run to the field and claimed they did it all.

he claimed they did it to themselves

If you don't believe what I shared is true, address that directly. But seeing my post sitting at 1 point and [flagged] after 2 hours is not OK. Just as DJT can't flag away his issues, you shouldn't be able to do so on HN.

One of the things I've loved most about HN is that it was real — grounded in observability, empirical evidence, not bias or feelings. I really hope that what happened to my post is not the beginning or a continuance of the end for that ethos.

1. I could have hidden my identify behind a throwaway. I did not feel that would be appropriate when making this calim.

2. I am not looking for anything, literally at all. Any follow ups for blogs; anything that would benefit I will not answer.

3. This is NOT a new account, I am very easy to find; I am 6'1 140lbs

I was working for a company called NationBuilder and I had the opportunity to go on a work trip. Outside of a talk he had just given I was waiting for my ride and I looked over like...damn thats the speaker. I wanted to say Hi; he damn near flagged down the police. I apologized and just decided to move on.

Note: It was in Reno, and no I don't want to go into details; the others are not hard to find because I happened upon them via blog posts so i'm sure if someone with the accumen of RF wants to know, he will find.

I have heard similar stores from several people in the years since. I AM NOT CALLING THIS PERSON RACIST. I am saying; he is observably scared of black people and that is not someone I want making descions about how the world moves foward.

FYI: I am by far not the only one to have experienced this and it 100% impacts hiring and other decisions at OpenAI.

Their value prop had to be strong enough to get past YC, past the other founders in the batch, past due diligence. Given that, I'm no longer comfortable casting "fraud" as a clean binary.

To be clear — I do genuinely believe they are a fraudulent company that lied and deserved to be removed. But introspectively, I have to sit with the fact that the space between "working around dumb regulations" and "outright fraud" is murkier than we'd like to admit.

Comments URL: https://news.ycombinator.com/item?id=47594828

Points: 2

# Comments: 0

When there's higher violence and lower property values in a Black neighborhood, people like OP are quick to blame Black culture. But when the "Cognitive Dark Forest" emerges from a community that shares its own common characteristics, suddenly collective accountability no longer applies.

When discussing violence in the Black community, it's "cultural." But when the subject turns to financial crimes or exploitation — where the per-capita ratios tell their own story — proportionality and population-to-crime-rate analysis mysteriously stop mattering.

It's difficult to take the "Cognitive Dark Forest" seriously as an existential concern when the people raising the alarm are so selectively offended. The crisis only becomes real when their innovations, their livelihoods, and their moats are threatened. Everyone else was supposed to just adapt.

The "Cognitive Dark Forest" is and will be continued to be perpetuated by "them" and if you really cared about the issue you would have addressed them.

One issue I see in your messaging is that construction people usually do not speak like "completely offline, no cloud, no accounts, no subscription". When doing work for SME's one of the things I enjoy most is learning systems, tool chains, nomenclature etc... The above reads like a dev harvesting for leads.