Humidity is the key, Finnish style sauna is low humidity+ high temperature (85-115C is OK i think), while Russian banya-style is low temperature (60-80C with high humidity). Both of them produce about the same load on a human
My problem with turkish style hammam is that unless it's extremely well maintained it often smells of mold. When I went to some nice hammams in turkey, I didn't have that problem but outside of turkey, it's often unbearable.
That's interesting. I don't have much the habit of doing sauna, as you can likely tell, so I might have tried only high humidity saunas. I'll give it a try one day with low humidity if I find one.
73°C isn't unusual. I checked out what's source for the Wikipedia article that says it's 80 to 110°C. Oddly it's a Chicago Tribune article from 1970. I don't think I ever visited a 110°C sauna.
110C is not that unusual in the Nordics (although way above average, it's for tougher sauna goers). I've been in one. Not most people's cup of tea though, the experience is comparable to the opposite of a long cold plunge.
A dry sauna sounds terminally boring. The point of Finnish saunas is that they are dry and hot, but you can adjust the pain...experience, I mean, by throwing water on the rocks at intervals of your choice.
This temperature cheating is one of the things I see very often in Gyms & public places:
They announce with "fin sauna 90°", and then its only 80 or 82,so stealing some performance :-D
I was in a 110C sauna for 20 minutes today. Plus 15 minutes in a 70C one (hybrid infrared sauna). Max is 30 minutes at once at 70C. It does take some getting used to.
With arch+hyprland I hit 5GiB for a zen browser instance with 15+ tabs and a kitty instance with 15+ windows across 5 tabs, with codex and vim running.
If ram is a problem there's always alternatives. The impediment is always having to rethink your workflow or adopting someone else's opinion.
Instead of Keycloak, I would recommend giving Kanidm a try: It's much more lightweight and covers most of what you usually need (one notable exception being SAML).
I was a long time k8s skeptical, but I think it's solid now. If there's good support for keycloak for k8s with support for backups I wouldn't think twice.
Not sure the state of keycloak now, but it was a lot of work to manage keycloak configs with the IaC pipeline. That could have gotten better now, but I think having access to the data is important because migration might not be trivial if for instance a provider starts acting up.
But isn't the one dimensional tokens a reflex of high dimensional space? What you see is "sure let's take a look at that" but behind the curtains it's actually an indication that it's searching a very specific latent space which might be radically different if those tokens didn't exist. Or not. In any case, you can't just make that claim and isolate those two processes. They might be totally unrelated but they also might be tightly interconnected.
I assume in practice, filler words do nothing of value. When words add or mean nothing (their weights are basically 0 in relation to the subject), I don't see why they'd affect what the model outputs (except cause more filler words)?
Politeness have impact (https://arxiv.org/abs/2402.14531) so I wouldn't be too fast to make any kind of claim with a technology we don't know exactly how it works.
Actually you'd trivially disprove that claim if you're starting from mechanistic knowledge of how orbits work, like how we have mechanistic knowledge of how LLMs work.
You have empirical observations, like replicating a fixed set of inner layers to make it think longer, or that you seem to have encode and decode layers. But exactly why those layers are the way they are, how they come together for emergent behaviour... Do we have mechanistic knowledge of that?
I think we've *only* got the mechanism, not the implications.
Compare with fluid dynamics; it's not hard to write down the Navier–Stokes equations, but there's a million dollars available to the first person who can prove or give a counter-example of the following statement:
In three space dimensions and time, given an initial velocity field, there exists a vector velocity and a scalar pressure field, which are both smooth and globally defined, that solve the Navier–Stokes equations.
Though the above exchange felt a tiny bit snarky, I think the conversation did get more interesting as it went on. I genuinely think both people could probably gain by talking more -- or at least figuring out a way to move fast the surface level differences. Yes, humans designed LLMs. But this doesn't mean we understand their implications even at this (relatively simple) level.
I don't think this is taking it as far as it can go.
Everything should live in the repo. Code and docs yes. But also the planning files, epics, work items, architectural documentation and decisions. Here is a small example of my Linux system doc: https://github.com/gchamon/archie/tree/main/docs
And you don't need to reinvent the wheel. Code docs can like either right next to it in the readme or in docs/ if it's too big for a single file or the context spams multiple modules. ADRs live in docs/architecture/decisions. Epics and Workitems can also live in the docs.
Everything is for agents and everything is for humans, unless put in AGENTS.md and docs/agents or something similar, and even those are for human too.
In a nutshell, put everything in the repo, reuse standards as much as possible, the idea being it's likely the structure is already embedded in the model, and always review documentation changes.
I think it's like for me moving to using Instagram in the browser instead of launching the app. Doesn't fight my addiction to doom scrolling directly, but it makes it awkward enough so that the small imposed friction reduces the urge with time.
If it works for you as it worked for me it's fine, but don't feel frustrated if you end up recreating bad habits in these kinds of setup. They can work but they don't really treat the root cause of the addiction.
But the system is proprietary, it's not yours. I don't get it with apple users. It's fine to purchase apple devices, they are gorgeous, well built, stellar performance and the UI is nice. But they never promised to keep an open system and to give you access, so why expect it? Even if you had an specific liberty with the system before, you were never entitled to that feature you lost after an update because the system just isn't yours.
> There are two obvious approaches: start with lots of guardrails, or start with very few and learn what the models actually do.
> We chose the second because we didn’t want to overfit our assumptions.
> Some of it went better than expected.
> But they also broke in very unexpected ways, sometimes spectacularly.
You clearly missed the whole point of the article, which is to experiment with agents and explore the limits of having them run wild.
Efficient use of tokens and which tasks to delegate is secondary to the experiment. Optimizing these is in any case premature if you don't understand the limits of the models.
I think you completely missed the point - they built a product purely using agents and deployed it to production for others to use. Read what the product actually does first.
Why shouldn't they ship it to production if the experiment was a success? You say the only way to code is to "learn to appropriate the correct usage of algorithms and AI" which for you is to code a generator and only use "dumb" generators to produce code, which is fine, but they just showed that for 20 bucks and a few minutes you can get very far, so their evidence is just stronger than yours.
What evidence? There is 0 evidence. It's deployed to production, but that doesn't mean it works fine or is free of bugs - which is exactly my point and why you use algorithms for these types of things. They're testable, repeatable and scalable.
And also because apparently "nobody has ever been fired for choosing Microsoft", which is something that should start happening more often if you ask me
Woah, that seems like a lot for me. I can usually stand maybe 60ºC for like 10 maybe 15 min. I don't think I'd be able to stand 30 min under 73ºC.
reply