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

A side note is that nvim does incorporate things upstream natively.

I believe "gcc", a la tpope, is now default part of nvim (not as a plugin just as a thing).

Also treesitter.


Even package management is built natively into nvim now (and nvim's implementation is supposedly inspired by mini's).

Same author iirc

> I'm just a regular dev, without any kind of attachment to Jane Street, in fact i'm a little wary of them, and would subscribe to Evan Czaplicki's understanding of the faustian bargain working there would involve.

Mind going more into this, since you named a specific person? Did he make some prominent statement?


Jane Street's whole business is front running and arbitrage. They make money in a way that's extremely unfair to the point a whole new stock exchange (IEX) was designed to stop it, hurts both retail and institutional investors, and provides no useful service to society.

All they've ever done for anyone but themselves is work on the OCaml ecosystem which is not exactly the biggest or most innovative.


Man, HN has really started to decline. How exactly is Jane Street front running?

Reacting faster than others to markets is not front running. Seeing someone else's order and putting yours in front of it is front running. Traders on the NYSE Floor would hear a big order is coming in, and put their ticket in front of it- that is front running.

In a computerized world, front running has essentially ceased to exist- the last major enforcement actions I could find are from 2003, when there was still quite a bit of trading done not by pushing buttons on a computer but by making phone calls. Trading on inside information is alive and well, but this is not what JS does at all, that's the realm of hedge funds that have traders that still click on buttons.

I never worked at JS, but have worked in MM.


Pay for Order Flow is not literally frontrunning and you can try to justify it as not literally fleecing retail investors but would just be dishonest.

I'm sure they make money using HFT strategies but I wouldn't say it's their whole business or even their largest profit center. They put on trades and take on risk in addition to having a market-making business. They're not the reason IEX was formed. Maybe they were a factor but idk.

> Jane Street's whole business is front running and arbitrage.

I instinctively hear a reply to this in Jeremy Irons’s voice, “We are selling to willing buyers at the current fair market price”.


They provide liquidity!

It's a case study here: https://www.youtube.com/watch?v=sl1UQXgtepE I think the critique is implicit so YMMV. Just my interpretation.

YOu're selling your soul for $ and telling yourself that you aren't extracting rents or directly harming retail investors but are just helping the market with price discovery.

Is there a source for this? I googled it and this was claimed to be satire

Not really tbf, it's an extrapolation. As far as I am concerned, there is growing concern, with Anthropic coordinated lobbying against open weights. The point I am making is we need strong Hardware actors consistently standing for hardware neutrality and integrity. Historically, NVIDIA has been very pro open weights. But at the same time, in-silicon guardrails seem to become the new thing: https://news.ycombinator.com/item?id=49879883

Makes you wonder what would happen if the US ever wanted to enforce "certified US AI" on its chips


I'm curious how effective something like that would really be.

Nvidia can certainly put up some guardrails that will sound impressive to the dinosaurs in Congress, but can it really stop China from using those GPUs anyway?

This is a simplification, but is this not in some way similar to how the Switch 1 has hacked due to a design flaw on Nvidia's Tegra chip?

In the same way, even if their GPUs get this protection, wouldn't a state actor figure out how to get around it?


The entire article that thread references never mentions any hardware, at all, only their software platform. The "Watchdog" referred to in the (editorialized?) title is literally never mentioned, and it is not really a piece of "hardware", it is a piece of _software_ called "Sentry" that (clarified in the Nvidia blog) they run on Bluefield-4 hardware (the "chip") which is their network card. It only exists in $10mil+ clusters. Just an awful and misleading article and headline, and an awful thread filled with people who know nothing about how hardware works arguing over nothing.

Most of the hardware functionality from a GPU that models need is generic and used every time you use a GPU for any compute. You can implement language models in a GLSL shader; modern GPUs have been programmable for 20 years, that's the point of their existence. The best thing they could do is just literally nerf the consumer hardware substantially to make tokens unusable or stop selling consumer hardware. They don't need spyware to do that. It would also make their products dogshit.

More broadly, "Compute" is basically a resource that our economy relies on to a large degree, it has been that way for decades. And it turns out it's very bad when you just destroy the thing your economy relies on. And doing so doesn't solve problems like "Current computers exist today" or "Other people make computers". So who gives a shit about Jensen? Comments here talking about "taking him out"? What's actually more likely is just that really powerful people today are often corrupt and also really fucking stupid. He played a stupid game and lost.

None of this stuff is putting the current open weights genie back in the bottle IMO. Amodei using RoboStormtrooperGruppen9000 bots and going door to door to take peoples GPUs at gunpoint is a more plausible method of doing so.


Yeah, even setting aside the impact and implication on the AI-tech-driven stock market of such a mandate, I don't think it's technically doable to implement either. What if you modified/wrote your own inference engine which uses a somehow slightly modified version of model weights, how would you want to detect (at a firmware level) that compute to be an "unsigned" model over, I don't know, a video game?

Open-weight models, at an absolute minimum the ones that already exist, are here to stay, and will continue ([0]) to get drastically cheaper to run.

[0] I shall continue to reference one my favorite articles in recent months: https://jyn.dev/tokens-too-cheap-to-meter/


Basic problems such as change over time and interoperability are reasons why formal methods break.

Formal specification presumes a priori a great deal of control over the world in general, especially the more you move away "from nature" and more towards what a human might want to instantiate themselves. Formal methods often try not to deal with things they don't "understand", or try to abstract the parts that "don't matter" away. But inevitably it changes, and the repsonse to that change is to subsume more and more into your own ontology so you understand the nuances. Or you simply control all interoperable endpoints, and so the nuances vanish away, because you're in control.

This is great when the battle is against nature. "Nature", for all its ambiguities, we generally can assume to be stable. What about with human society?

In some sense, there might not be much to formal methods if you could solve the underlying capability of getting everybody to agree on the semantics in the first place, and have no desire for change at all. But that is an absurdly strong requirement. Formal methods is constantly a race to try and find good abstractions and then breaking them.

I don't think there really is a way out of this, but I'm not that creative :/


I don't think a simple prompt will exceed any kind of memory system. Nothing fancy, but you need the agent to tune to your quirks; I find that a prompt to try and distill everything a priori is too annoying.

Been just using less of chat recently and using more codex for learning - a direct shared interactable terminal is nice, it's still relatively cheap, and importantly you can log specific thoughts to say a git repo for memory.

It's not attempting to be some kind of knowledge ontology graph like many businesses want - you just want it to understand the specific quirks of your specific learning path, which may not be formalizable or worth be spending time to step back and formalize every little thing.


Interesting, will take a look

mDNS my goat. I have a lot of spare machines and I do a lot of various experiments on said machines, some of which could brick the system even on NixOS. I now have a custom usb installer over mDNS and reflashing any compute I have from scratch for the express purpose of destroying it in a few days hasn't been easier

My job requires thinking about things like this (well, more senior and experienced people actually drive the final decisions, but I still need to understand them). Seeing the level of rigor that went into the development of a stable C abi for this is insane, it hits on pretty much all the checklists (and more!) that we have internally.

Add me to the list lmao. It seems like this is pretty popular among folks


Maybe just teach him more math?

Honestly the school system killed my passion for math, before I knew about more advanced math. Getting a headstart in say basic abstract algebra (but without too much abstraction - just some basic structures, to then apply to real world schemes) might be interesting. Polynomial structures also have connections to physics and SDR, and RF/crypto go hand in hand (comms)


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

Search: