lagrangian
5mo ago
Interesting, I like even plain cruise control. Gets tiresome on my ankle otherwise. But adaptive is a whole nother level. Now you don't have to speed up and slow down as people in front of you do things - you are really just setting a max speed. Super great in stop and go traffic.
lagrangian
5mo ago
Are you saying you consider moldy bread edible? I throw away the whole loaf, and would look at anyone doing otherwise strangely. I'm wondering if this is a geographic/cultural norm - I'm from the US, for context.
For some reason, growing up, for cheese we would cut off the moldy part plus a couple inches, but I've mostly chalked that up to "grew up less well off than I am now" + "Jewish cheapness".
lagrangian
5mo ago
CFPB:
The Consumer Financial Protection Bureau (CFPB) helps consumers by providing educational materials and accepting complaints. It supervises banks, lenders, and large non-bank entities, such as credit reporting agencies and debt collection companies.
(Not a response, just adding context)
lagrangian
5mo ago
lagrangian
5mo ago
Interesting, I was hoping for more detail than that from them. Thanks for digging it up.
This is a delicate moment. It’s a new day and yet so many are on edge. Our politics are inflamed, and social media only adds to the anxiety.
It seems that @elonmusk made an awkward gesture in a moment of enthusiasm, not a Nazi salute, but again, we appreciate that people are on edge.
In this moment, all sides should give one another a bit of grace, perhaps even the benefit of the doubt, and take a breath. This is a new beginning. Let’s hope for healing and work toward unity in the months and years ahead.
lagrangian
5mo ago
especially with the ADL coming out to defend him.
Can you point me to what they actually said? I can find lots of secondary sources claiming they did, but none containing a quote, and nothing on adl.org.
It seems plausible to me. Maybe something like "calling people who aren't killing Jews Nazis is like, kind of in poor taste, please stop."
lagrangian
5mo ago
Yeah, the closer I look, the less I am impressed with its comment.
lagrangian
5mo ago
[Meta] I gave gemini the whole contents of this thread, and asked it to select a comment to disagree with. You "win":
I disagree that AI inherently prevents a "conversation between people." It's a tool, like any other. Sure, it can be used to disengage, but it can also be used to sharpen arguments, explore different angles, even generate thought-provoking counterpoints. The problem isn't the AI itself, it's how people choose to use it. Just like a human can be stubborn and refuse to change their mind, so can someone using AI badly. But that doesn't mean all human interaction is invalid, right? We shouldn't throw the baby out with the bathwater. A well-integrated AI contribution can actually enhance a discussion by bringing in perspectives or information a single human might not have considered.
So how do I feel about this argument? It's coherent, representative of the pro AI arguments here...and cripples the ability to draw signal about what people actually think. My post is at least tolerable under "use/mention" but, ugh, I feel dirty. qa!
lagrangian
5mo ago
In debates over whether it's ok to use AI for X, it's helpful to ask "instead of asking an AI to do it, what if I asked another human to do it? Would that be ok?"
I like the metaphor, but I think it misses the key detail that AI effort has a trivial effort floor. Your "copying from substack" metaphor is spot on though.
lagrangian
5mo ago
lagrangian
5mo ago
If you mean behind the scenes, sure*. If you mean "and then quote its facts/figures," no. I consider "AI says $STATISTIC" to be at most twice as accurate and at least twice as irritating as "my ballpark guess is", while being significantly dishonest. It's just forcing others to do your work for you.
*: Even researching is on shaky ground. "Question -> AI -> paraphrase answer" is marginally better than piping AI into a textbox. "Question -> AI -> check original sources -> synthesize your own argument" can be done well. I personally don't find it more useful than "Question -> search results -> check original sources if those weren't already -> synthesize your own argument", but concede that I am a luddite. (muh vimmmmm)
lagrangian
5mo ago
the onus is on the user to put their own time and effort into vetting and fact checking it
i am concerned that in practice it's going to fall heavily onto other users and the mods, rather than OP.
i think we can have ~all the ai you actually want with a policy of "no ai, except for short things where you used it so well/minimally that we can't tell and it's more like a spell checker than outsourcing"
lagrangian
5mo ago
Can you make any argument in defense of your apparently instinctual reactions?
Yes. In short, AI makes it too easy to be low quality given the finite energy and speed of our mods.
Metaphor:
Suppose you're a camp counselor (mod) and find that your campers (posters) sometimes smell bad (make bad posts). Ideally, you could just post a sign that says "you must not smell bad" (don't write bad posts) and give people total freedom, simply punishing those who break the rule.
In practice, you need stricter rules about the common causes of the smell, like "shower daily" or "flush the toilet" (don't use AI) - even though, sure, some people might not smell bad for a couple days (use AI well) or might like the smell of shit (not hate AI generated prose like I absolutely freely admit that I do). It's just not realistic to expect campers (us) to have good hygiene (use AI responsibly sufficiently often).
lagrangian
5mo ago
Could we hear from a mod who wants an AI policy even as permissive as "quoted just like any other speaker"?
My two cents:
How AI generated content can be displayed. (off site links only, or quoted just like any other speaker)
off site links only, other than very short quotes not making up the bulk of a comment, and even that I kinda hate
What AI usage implies for the conversation.
the end of my interest in a thread and a sharp drop in my respect for the user
Whether a specific rule change is needed to make our new understanding clear.
yes, please. otherwise, it's far too easy to spam, lowering quality and increasing moderator effort.
Bottom line, I think we need to discourage AI heavily. Otherwise, long form content - the hallmark of much of the best content here - is immediately suspicious, and I am likely to skip it.
lagrangian
5mo ago
I feel like if went through a year long dry spell I could, more-or-less, shag anything.
Amen. Been there, shagged that.
lagrangian
5mo ago
Like if I was in my 60's I'd stay, but not in my 40's.
I think this would be a more compelling argument if you were in your 20s, or mayyyybe your 30s. Unless she's super hideous, I have to think you could enjoy fucking her. It sounds like you've done it enough times that revealed preferences say you like it plenty.
If you're super hot and expect to do really well on the dating market, and enjoy the idea of starting over with a new person, then...my advice is still "you're old, the ship sailed, make the best of it." But I guess it'd be more reasonable in that circumstance to try again, at least.
lagrangian
5mo ago
I haven't actually heard you say why you feel you haven't had sex, explicitly anyway. Could you just choose to do it?
I feel like there is a significant range of attractiveness that isn't sufficient to be all that exciting from a strictly visual standpoint, but is entirely sufficient to want to sleep with someone you love.
I don't think it's as obvious as others seem to that not sleeping together for one year is immediately a problem. I would happily take the least attractive of my exs' bodies with my favorite ex's personality over the reverse.
lagrangian
5mo ago
Hm, interesting. Care to share the code, via post or DM?
You say "CAS" but also "arithmetic". What's the story there? If it's heavy symbolics, I have one set of advice; numerics, another.
lagrangian
5mo ago
Premature optimization is the root of all evil.
Especially if your problem parallelizes, the answer may be "whatever is easiest for you to write in" plus one or more honking cloud VMs. Consider also writing it in whatever is easiest for you, then having an LLM port it to something else.
If you don't need speed for iteratively developing, it really may be best to not care about speed, then let the thing run for a few weeks.
lagrangian
5mo ago
Electrolyte pills will have more quantity and types of electrolytes than a gatorade, without the sugar. I have taken this brand.
lagrangian
6mo ago
I think some variety of "I'm an idiot" is more likely. I don't see any images. If they're included as hyperlinks, they're not loaded until you click (I think, and ~confirmed by dev tools)
Attached a screenshot of the resource usage breakdown. Largest element is 402kb for the banner, compared to 215kb (fifth place) for what I think is the actual comments (compare to my 414kb estimate - not bad).
Some of the overestimate is from my extensions. Filtering those out, I see 2.0MB (2.3 uncompressed). 1.15 of that is fonts (unclear to me why that needs reloading each time - presumably this could be optimized out.)
Off the top of my head, $50.
-
Fermi estimate:
Multiply the following:- 207 "report" ctrl+f matches = comments
- 5 lines/comment
- 20 words/line
- 5 characters/word
- 4 bytes/character
- = 414kb
-
Comparing to dev tools, which shows 5.3MB, a factor of 10 I can't account for.
- I'm a backend dev...what's an order of magnitude between friends.
- Using that figure and 24k thread views on the culture war thread so far this week:
- = 127gb/week
- = 210 kilobytes/second
-
Let's assume we want to serve a peak traffic of 10x average and don't care enough to set something that automatically scales up and down:
- = 2.1 megabytes/second
-
This is... jack shit.
- A 3.5" floppy disk can do 100 kbps, filling the entire 1.44mb in 14.4 seconds.
- I think it could probably be served off a Raspberry Pi.
- If the vyvanse hadn't worn off, I could probably calculate how many threads/Ghz/etc are needed, but I'm pretty comfortable saying "one of any shitty processor can handle this load"
-
Google cloud charges for egress:
- Checks notes:
- $0 for up to 200GB/month, then $0.11/GB up to 2TB.
- So
(127 * 4 - 200) * 0.11 = $34. - I am not actually sure if serving traffic is "egress". Best guess: no.
- (This is why startups shouldn't hire FAANG engineers.)
- Checks notes:
-
Worst case:
- Something like $34 for egress and $20 for the VM itself.
- Pretty close to my first number!
-
@ZorbaTHut, how'd I do? And would you be willing to share what % of costs you've had donated, vs paid out of pocket? You really shouldn't have to be paying yourself.
- Edit: the patreon is at $140/month, so it looks like this site may be slightly profitable (ignoring the enormous value of the free labor). Nice!
That was all without chatgpt, but here's a transcript from my talking to it afterwards. I think it looks reasonable until maybe the end when I ask about vpn costs. Still comes out to ~$50. It did a decent job analyzing the amount of cpu used (which I skipped in the "jack shit" section).
lagrangian
6mo ago
So in comparing, say, this site to Reddit, there's probably some complex code for managing the orders of magnitude greater traffic that themotte just doesn't worry about?
Right. Zorba pays for the site out of pocket, but that is not scalable. The site occasionally goes down - we even lost most of a day of posts not too long ago. That's no big deal at our scale - just ssh in, figure out the bug, deploy something manually, etc.
But at e.g. Google scale, it's $500k/minute of gross revenue on the line in an outage, to say nothing of customers leaving you permanently over the headache. Fractions of a percent of optimization are worth big bucks. Compliance headaches are real. Hardware failures are a continual certainty.
Read about the brilliance behind Spanner, the database a huge amount of Google is built on: their own internet cables, because why choose between C[onsistency] and A[vailability] when you can just not have network P[artitions]?
You need an incredible degree of fault tolerance in large systems. If n pieces in a pipeline each have a probability p of working, the system has p^n - exponential decay.
Plenty of it is feature bloat, that said. You really can serve an astonishing amount of traffic from a single cheap server.
+1. That sounds quite excessive to me.
See e.g. https://www.niaaa.nih.gov/alcohols-effects-health/alcohol-drinking-patterns which defines "heavy drinking" as >=5 drinks on ANY day, or >=15/week. You're at 35-56/week. If you drink it very slowly, it's a little less concerning, but I think you're way past the gray area. A doctor friend told me they start asking questions (but there may not be a problem) at 14 drinks/week.
Or, in calories, that's 100/shot = 3500-5600 calories/week = 1-1.6lb/week. If you're drinking beer or fruity cocktails, multiply by 1.5-3. I'd even be concerned about someone having that many calories of icecream per week. At the same weight, you could replace with healthier foods and your body would be running on much better fuel.
Or, in dollars, even if you're drinking fairly cheap booze at $1/shot, that's $35-56/week = $1820-$2912/year. Might or might not be a lot of money depending on your job, but that feels nontrivial even from a FAANG salary. Especially since, if you're anything like me, a dollar spent on booze often means several more spent on munchies.
Yup. It really is that simple. I hold VFFSX, although there are a number of similar low-to-zero fee large-cap US stock indices that are all "S&P500".
Even if you think you could beat the market, you have to decide if it's worth the stress/effort. I think it's very much not. Surely, if optimizing for dollars, your effort is better put into advancing your medical career. Similarly, if you find yourself cleaning toilets at home, either a) capitalism is fundamentally broken or b) you should hire a housekeeper.
Some mix of VFFSX + safer or anti/un correlated asset classes makes sense if you are either particularly risk averse or particularly concerned with timescales shorter than 10-15 years. But, this leaves expected value gains on the table, and requires more thinking, so I'd advise mildly against. See e.g. Butterfly Portfolio.
Beyond that, two notes:
More options
Context Copy link