Check out signoz: https://github.com/SigNoz/signoz

OSS o11y platform built on clickhouse & otel.

Someone can likely give you a more technical explanation - but to give an anecdotal example, my parents have German Shepherds which are grazers. They put a few cups of food in their bowl in the morning, and the dogs will eat throughout the day as they're hungry. They like treats, sure, but I wouldn't call them food motivated. My dogs are lab mixes, and if food is out they'll sniff it out and immediately scarf it down. When one was younger, if we left food on the counter or table while he was home alone, he would jump onto the counter and eat it. I would call my dog very food motivated.

How could you know if the German Shepherds are food motivated if they literally have food available to them whenever they want? I would guess if they were fed at discrete times of the day they would start looking at the owners for food.

That is a difference between "starving" and "being food motivated".

Think about humans: all humans need calories, but some people just want to drink Soylent and other people will spend hours creating elaborate meals. Some people will barely eat because they know they need to, while others will eat until they are sick if they don't stop themselves. Some people will walk miles for the best ice cream, while others will walk miles to see their favorite band.

Food drive is a thing in humans too, and it's something that's largely genetic -- and it's something that you can modify with anorectic drugs like GLP-1 RAs and stimulants. How much you eat is a combination of intrinsic food drive and trait conscientiousness.

Everyone knows someone who will eat like 3 potato chips and move onto something else, while others will sit and mow down the whole bag. Taking a GLP-1 will basically just convert you from one kind to the other in a dose-dependent way.

The same kind of variability is present in dogs - but without the conscientiousness axis.

All the training in the world goes out the window when a GSD can graze.

But if there's something tasty going down like diced chicken breast or some steak, they'll show you they remember every learned command perfectly.

Even mentioning their favourite food gets their attention as some words are worth learning and remembering.

I had a husky-lab mix. He was not food-motivated at all. We did _not_ leave food for him all day long, because the other doggo, purebred lab, would eat the food designated for the husky if we did.

So we fed them both at the same time, twice a day, and had to watch if the husky would eat or not for that meal, and remove the food bowl if he wasn't hungry, so the lab wouldn't scarf it down.

This is an n=1 anecdote that your guess was the opposite of my experience. That dog just didn't care about food. He did however love eating ice cubes. I never tried just a bowl of ice cubes.

Supposedly IQ test have a slight positive correlation with workplace performance. I could see leetcode skills(or, put more broadly, algorithms knowledge) having a slight positive correlation with job performance, but as everyone knows it doesn't guarantee someone is a good or productive developer.

The flaw with the metaphor is that you can study for leetcode, whereas supposedly(can't find a good source on this) you can't really study for an IQ test. I had a weak comp sci background when I started doing leetcode interviews and would struggle with questions. I spent several months studying CLRS, cracking the coding interview, etc., and now I would say I am fine at most problems up to medium complexity.

Studying leetcode and IQ both correlate with better performance in leetcode.

There's no raw data to back either statement up, but both are very likely to be true via common sense.

I believe there were rumors they are developing a commercial model: e.g. https://www.ft.com/content/01fd640e-0c6b-4542-b82b-20afb203f...

Plus customer facing employees aren't just the people in the branches, there's a massive volume of phone/email/etc support any large credit card issuer receives on a daily basis.

After 5 years of this law being in effect, will the numbers balance back out to pre section 174? That is, does that deduction from year Y carry over into year Y+1, Y+2, etc.?

I mean, this probably still matters a lot for startups given their shorter lifetimes, but it seems any large company(I'm thinking of, e.g. Apple, who has plenty of cash on hand) that's been around for a while could just wait it out? I am not familiar with corporate tax law and how deals are structured, but could you also defer revenue in the same way to offset(e.g. customers with a 5 year contract paying progressively more but keeping the same total $ amount to sync with your deductions)?

It'd be neat if the software was opensource and you could point this device at your own LLM, then at least you'd have some assurance it won't be e-waste if the company shuts down.

But I don't see a world where you can feasibly support the required infra of providing LLMs for your users with only hardware sales as a revenue stream. Their website specifically calls out that there is no subscription fee, so either they'll introduce one later or something else will need to be the other source of revenue.

You would think that, but it's rarely true. Getting issued a $10 USB drive could take a ton of paperwork and red tape, and eventually end with a "yeah that use case seems inconvenient but not worth us doing stuff". When you have one sitting on your desk or in your pocket already, that's a tempting one

Honestly in my experience every company I've worked at has had some kind of explicit cybersecurity policy or training telling us not to use USB drives, so from a CYA perspective I think I'd be very hesitant to plug one in to a work laptop at all. But I have only worked at massive corporations, so my experience is limited

Philadelphia metro area has very weak wages compared to most other east coast cities for tech. We don't have much of a FAANG or tech company presence here to bid up market salaries(outside of a couple of satellite offices in the suburbs). I can tell you from my network that with ~6-7+ YoE you can earn about ~180-210k counting base & bonus at most of the banks or financial firms as a new hire at the VP or tech lead level.

I believe teamspeak is also pretty popular in the milsim community due to its support for mods in games like Arma 3

Oh yeah, ACRE2, which simulates real-life radios to the point of tracing radio signal paths, accounting for power output, terrain, buildings, antenna plugged into radio, antenna radiation pattern and other factors. Absolutely insane mod.