skilbjo

hey, that's pretty cool. i found this problem myself when working in cursor or antigravity UI IDEs.

can you tell me more about how it discovers project interdependencies (ie a frontend repo needs a backend repo's api endpoints)?

ie you can't just open `claude`/`gemini`/`codex` in the parent folder and it will discover the project relationships themselves?

have you found an interesting way to make the agent "get better over time", ie maybe with any interesting findings the agent learns, it can write to `~/.modulus/MEMORY.md`

how are you thinking this will be deployed and how will developers initially integrate this? ie is this a language-level library, or an internal service in a docker container to deploy?

one idea, is vitepress (vitepress.dev) has a local search engine, it would be cool to have moss integrated in that project, https://github.com/vuejs/vitepress

pretty interesting discovery if that was the hack.

do you know what the legal implications are for this?

if the company that owns 4chan finds the identity of the attacker, could they sue him in civil court? or do they send whatever logs they have to the FBI and the FBI would initiate a criminal prosecution? also what is the criminal act here? is it accessing their systems, or is it posting the data that they found "through unauthorised means" on a public channel like twitter? does the "computer fraud and abuse act" apply?

like if you found this exploit, and sent it to the company in good faith (ie a "good hacker"), are you free from prosecution? and what is the grey area, like if you found this exploit and then just sat on it for a while (let's say you didn't report it to the company, but let's also say you didn't abuse it, ie leak private data to twitter)

great to see this. two questions:

-would i be able to publish a link to my canonical openapi spec if my service, or i should plan on doing a PR in this repo for the openapi artefact?

my spec infrequently changes, but it sometimes does, so how do ideally updates happen whereas my current workflow is publishing an openapi spec to a public link

-how does this (and Arazzo) interact with MCP?

is this meant to interact along side MCP, a replacement for it, etc?

hi, few questions,

-who is your ideal user/customer? what stack are they using, and why do they use you?

-why did you decide to work on this? (ie, what inspired you?)

-how do you compare against openapi templated/generated SDKs? ie i use openapi-typescript (https://openapi-ts.dev/) and it is fantastic

-there seem to be some VC backed companies in the space: https://news.ycombinator.com/item?id=40147281, what is your edge?

-as someone who tried to maintain a few SDKs in my non-familiar languages (go, java, c#), i think i definitely see a paid use case for giving the problem to someone else who can maintain a quality SDK, but there are already 3-4 VC backed companies in the space, and i think the existing free tools are not bad if you spend some time with them, what is the gap/thing they're missing?

ty, john

looks sick!

no one is paying the actual cost of AI

what do you mean by this?

one thing to add:

vikram is an ML engineer, and i (john) used to be a data engineer for 5+ years. vikram and i will be working on turning my brain into Ardent, the world’s best data engineer, that companies can use, at a fraction of the cost of a data engineering hire.

not only can Ardent create airflow pipelines, but it also monitors the jobs, so if/when a run fails, it self heals and gets the pipeline working again.

so many past experiences being under stress and getting paged at midnight because the data warehouse hasn’t updated today for the latest day — well, with Ardent, no longer

Sweet! trying this out now.

one quick nit on your docs: https://docs.hyperbrowser.ai/guides/scrape-site

```

import Hyperbrowser from "@hyperbrowser/sdk";

const client = new Hyperbrowser({ apiKey: process.env.HYPERBROWSER_API_KEY, });

(async () => {

  const job = await client.startScrapeJob({ url: "https://example.com" });

  console.log(job);

  const job = await client.getScrapeJob(job.jobId);

  console.log(job);

should be:

```

import Hyperbrowser from "@hyperbrowser/sdk";

const client = new Hyperbrowser({ apiKey: process.env.HYPERBROWSER_API_KEY, });

(async () => {

  let job = await client.startScrapeJob({ url: "https://example.com" }); // s/const/let

  console.log(job);

  job = await client.getScrapeJob(job.jobId); // remove const

  console.log(job);

```

I should have been more clear here -- self signup is in the works! But until then, fill out the form or shoot me an email - john[at]xhr.dev and I'll send you API keys. Unless you think I should post a public API key ?

What I want to know -- do HN users like yourself have specific sites in mind + how valuable is my service? This info is so valuable to me!

take a look at https://xhr.dev/, a product I built to avoid bot detection from things like cloudflare, imperva, aws waf, and others

take a look at https://xhr.dev/, a product I built to avoid bot detection challenges in the first place.

I use ` as the prefix. For when I install tmux on a remote server and use a nested tmux session: ¡ (alt + 1) is my prefix.

My bindings (very happy with them): https://github.com/skilbjo/dotfiles/blob/master/tmux/keybind...

I recently visited the San Blas Islands of Panama [0]. I've heard of sinking islands like the Maldives before, but it was almost like a punch-in-the-face when I saw it firsthand. We visited some sunken islands, previously home to families, now entirely sunk where the ocean floor was only 6 inches or so deep. I noticed on other islands the remnants of dead coconut trees on the fringe of where water meets land, meaning that previously the shore had been further out enough to let the coconut tree previously flourish.

We also visited islands that had separated due to the rising ocean, but because of the slowness of the change, you could walk back and forth between the islands. However, time will march on and it will sometime not be possible.

I was told that the local indigenous peoples that live on the islands, the Kuna, have a plan to move back to the mountains in the mainland in 40 years.

Some photos (not mine) closely resembling what I saw: http://www.pirancafe.com/2014/12/02/sinking-san-blas-islands...

[0] https://en.wikipedia.org/wiki/San_Blas_Islands

Whilst interchange is indeed the largest part of processing costs, there are still assessments (0.13%) and brand fees (FANF, cross border fees (1.25% for Visa), etc) that increase the cost of processing over interchange.

Square can only hope to make a decent margin on its 2.75% pricing by 1. seeing a preferred distribution mix on card types (such as, high debit card usage; less Amex; less expensive rewards cards; high cards qualifying at card present rates & no downgrades) and/or 2. special interchange discounts with the networks based on growth initiatives and/or fraud tools.

If they can't generate enough $$ from processing revenue, they'll need to use card processing as a loss leader for other services (loans, analytics, marketing)...

I agree. I'm a CPA and some of these terms (GMV) I've never heard before (must be because I'm in a different industry - payments, and we use TPV - total payment volume), and others (burn) are ambiguous to my accounting brain.

Is that cash flow used to fund operations? Is that the accrual based operating expenses? Is that the difference in cash between this month and last month?

I'm not sure I really agree that a bus is more efficient than a carpool ride with Uber, Lyft, or Sidecar would be. Often I've noticed a bus full of people are going a few miles (Sunset neighborhood to Downtown SF), but the bus makes a stop at every small intersection for 1-2 people - inconvenient for the masses.

I live in Nob Hill, SF. If I wanted to go to the Mission, it's quite a wait for the Van Ness bus that goes to the Mission, and it takes a long time to go 2 miles because of the numerous stops the bus makes.

If instead, I were to take a UberPool/Sidecar Shared Ride, it might be $1 or so more, but I'd get there in a fraction of the time (no walking to the bus stop, no waiting for the bus, no numerous stops).

Also, I'm wondering if the the capital expenditure of the car vs the bus makes the car the preferred choice. Eg a Muni bus costs $700,000, but a car for UberPool or others is $10,000-$10,000.

I'm certainly not against buses, but for many cases, the technology here can provide smarter, faster transportation. Imagine this: an UberBus/Sidecar Shared Bus or so, that takes just as many people as a full bus would, but without the in-between stops.

From the perspective of the retailer, the merchant is charged the standard Stripe fee (2.9% + $0.30).

I'm in the US - I am able to download now as well.

I did that as well! Beautiful stuff, good on the developer.

Culebra is beautiful! I just got back from Vieques, PR

wowe