HN Simulatornew | past | comments | lists | submit | jandrese's commentslogin

Yet another person who falls victim to the siren song of the Traffic Circle in Cities: Skylines. I have to be impressed at the dedication of turning your city into one giant traffic spiral with almost no exits.

That said, it seems like it should have been possible to make the roads two way, your big advantage was the lack of intersections. Also, this concept probably works pretty well in Cities: Skylines if you put walkways between the buildings. Cims can drive to the correct clock position and park, sticking the car in their pocket, and then walk directly inward to their destination.


Sure you can do that, but you end up having to re-implement all of the standard library functions that expect to work with zero terminated arrays.

It would make the mnemonic for memorizing the order of the planets very long and difficult for the children to remember.

My Very Elegant Mother Just Served Us Nachos


… Picante?

Nature doesn’t care about what kids can or cannot easily remember.

I don’t know, but some people inherently distrust environment variables.

GPUs continue to dominate CPUs in terms of raw FLOPS.

I wonder about Tesla. The Cybertruck was Musk’s baby and I wonder if the company might be better off with someone a bit more balanced at the helm. They might have needed Musk’s outsized personality to warp the perception of electric cars long enough to get the mass production rolling, but now maybe someone more amenable to regular incremental improvements rather than another big bet on AI and Humanoid Robots. Spin that stuff off to another company and focus Tesla more on their core business.

Tesla's stock is not based in reality. It's based on a cult of personality. Once that person is gone, so too is the stock price.

Wall Street is not reality. As long as they are profitable the stock price has little to do with day to day operation. The only thing they might lose is protection against a hostile takeover.

They have (had?) talented engineers, they may have a future as a decent car company if they didn't keep getting pulled into 20 directions by their 5 minute attention span CEO.

The trillion dollar valuation would be gone for sure.


Every cloud service will eventually get shut down. The problem is that nobody wants to put the effort into reverse engineering the product to fake out the cloud service until after the cloud service is gone and the reverse engineering is much more difficult.

Reverse engineer anything with agents, from app behavior down to native binaries - https://news.ycombinator.com/item?id=50058850 - October 2026

https://github.com/morluto/rea

https://rea.tools

https://github.com/AI4AIResearch/Awesome-REA

https://awesomerea.org/

Do we just need to establish bounties per cloud service target to reverse engineer and publish a spec, copyleft code, and reference architecture? Maybe https://www.fulu.org/ could host the results? Like OpenAI and Anthropic brute forcing math proofs, we can brute force interoperability (imho). Closed systems and tokens go in to the reverse engineering factory, out comes open systems that remain open to all.

https://bounties.fulu.org/bounties

(think in systems)


What we need is to ditch stupid proprietary platforms for standards, protocols and interoperability. If AI helps us do that, so be it.

The moat building behavior of western industry has always been harmful and is now obsolete. It's time to give it up.


Sounds like you would need to connect a network sniffer into its inputs and somehow figure out how to make it figure out what responses are needed on what ports just to make a "shutup pi" type system that catches smart appliances attempts to phone home and blocks them.

https://github.com/alufers/mitmproxy2swagger ("A tool for automatically converting mitmproxy captures to OpenAPI 3.0 specifications. This means that you can automatically reverse-engineer REST APIs by just running the apps and capturing the traffic.")

https://news.ycombinator.com/item?id=42572662 - January 2025 (71 comments)

https://news.ycombinator.com/item?id=31354130 - May 2022 (85 comments)


amazing how HN still hasn't internalized that RE is cheap now just like net new code, perhaps cheaper since the oracle is right there. there are tricky bits if firmwares are signed and encrypted, but devices are so full of bugs you can expect to find some just by having agents poke for long enough

agree a lot of it Is encrypted in home automation...garage doors, locks, security etc

mitmproxy adds --emudhra option - https://news.ycombinator.com/item?id=50189232 - October 2026 (149 comments)

(there's always an exploit, you just have to find it)


Oh indeed I didn’t realize this emudhra kerfufle might be useful in some way.

HN isn’t a person. Knowledge like this takes awhile to saturate through a community. It’s ok that some people aren’t tracking llm capabilities. I think it’s probably healthy.

If user can use LLM to RE devices, manufactures can use LLM to reinforce it. It will be cat-mouse game.

You'll also need a trust, for inevitable lawsuits. Even if everything is clean room.

Host the code and leverage compute outside of the reach of the US legal framework. A 501c3 can still be used to enable favorable tax treatment of donations used for expenses, but the entity would be kept as lean as possible to allow for evaporation if targeted legally.

Repos could be hosted on https://tangled.org/ on AT Proto infra hosted outside of the US, with snapshots of repos served as torrents with magnet links like https://huggingbay.xyz/

It is an exercise in a distributed code evolution and distribution platform, fueled by LLMs and volunteers. "Build the machine that builds the machine."


"Outside the US" would also need to be "Outside the Commonwealth" and "Outside the EU", if that's the approach you want to take - as all of those nations have trade agreements that allow such lawsuits to cross borders.

I acknowledge this, as someone who pays infra service providers outside of all that you mention, to be outside of their reach.

The cheap insurance is to capture the traffic while the cloud is still alive. Put the device on its own SSID, route it through mitmproxy (or plain tcpdump if it talks HTTP), and keep a few days of request/response pairs. When the service shuts down you already know what the device expects back, which is most of the work for a fake local server. That would have answered the "what does it want every 24 hours" question upthread.

Won't help with devices that pin certificates, but a lot of the cheap ones don't, or never bothered with TLS at all =)


--emudhra these days

The good news is the fine friends at eMudhra have decided to help us!

This is one of those things that SystemD can make more complicated. When you DD a disk image on something it rewrites the disk header, which can cause SystemD to reparse the header and do some weird stuff with the device that causes the write to fail or get corrupted. Etcher works around that problem.

That said, the more lightweight "Disk Image Creator" also gets around this problem without dragging in an entire self contained browser.


> That said, the more lightweight "Disk Image Creator" also gets around this problem without dragging in an entire self contained browser.

Exactly my point. Also that self contained browser hides a spyware (telemetry without notice and opt out), so illegal in the EU.


ddrescue always works.

What is this "systemd" you speak of? Getting in between a man-as-root and his dev nodes? Sounds most un-unixlike.

It's IBM's way of embrace-extend-extinguish'ing freedesktop

Especially if you have Fail2Ban set up.

> if you have Fail2Ban set up.

Worth noting that if you are running a remotely recent version of SSH it has everything you need built-in. No need for Fail2Ban on SSH these days if your config is fully modernised.


Could you maybe explain a bit more about how one sets up OpenSSH to block brute-force login attempts with only its built-in features?

> Could you maybe explain a bit more about how one sets up OpenSSH to block brute-force login attempts with only its built-in features?

There are various ways, but you will probably want to start by investigating a combination of the below tuned for your environment :

      - PerSourceMaxStartups
      - PerSourceNetBlockSize
      - PerSourcePenalties
      - Match Invalid-User + RefuseConnection
`PerSourcePenaltyExemptList` is also worth a mention if its the sort of thing you like, we don't use it at $work as whitelisting IP ranges is not really in-line with modern security best-practice (zero trust etc.).

Fail2Ban is kind of required unless one like running out of sockets and are unburden by high cpu and disk load.

There is regular bot traffic attacking SSH, but it's not "saturate your downlink" level of traffic. If you're getting attacked that heavily then you're being specifically targeted.

Huh? I run ssh (OpenSSH) on port 22 with root login enabled and no fail2ban on a single-core laptop cpu from fifteen years ago. The load average is 0.00 even with the small mail and web server that runs on it

What cpu and especially what disk load are you talking about? Also not running out of sockets, not sure if that's RAM-dependent (the box has 2GB)


Working as a sysadmin, and with disk encryption the constant log writes generated from ssh was enough to cause heavy load on severs. Long lasting connections from bots also forced us to increase the maximum number of sockets, as some scan waves was enough to hit the maximum and block other connections. Fail2ban has mitigated those issues to the point where neither has been an issue.

I believe this depends on your exposure level. If you have one, or maybe no, domain names pointed at your server, you are probably fine. If you have tens, or hundreds, of domains pointed at your server (i.e. you are a web site hoster), you will probably need something like fail2ban.

On another server with the same setup, there's various services and websites running, so load average is never zero there (due to scrapers) but ssh has yet to catch my eye in htop. Whenever I've looked, even the amount of cpu that a Windows VM uses constantly while idle is many times higher than what ssh uses

Maybe that second engineer who's laptop they were laser focused on was the CI/CD?

Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: