bjorney

Does Google Cloud not count as “own hardware” for google?

That’s why the bars are so different. The “cloud” price is MSRP

That’s not really machine learning though. If you wanted to go way back, AI research goes back to implementations of hebbian learning in computer science back in the 1950s as a way of emulating human neurons. I was merely pointing out that AI was a computer science “dead end” until restricted Boltzmann machines were revisited by Hinton et al back in 2008 or so, and that 99% of the growth in the field has happened since the early 2010s when we reached a turning point where deep learning models could actually outperform classical statistical models like regression and random forests

in the coming decades

Given that in the past 15 years we went from “solving regression problems a little bit better than linear models some of the time” to what we have now, it’s not unfounded to think 15 years from now people could be giving LLMs access to code execution environments

Pagers operate at a lower broadcast frequency than cell phones. Longer wavelengths (low frequency) are less impeded by walls and interference.

This is only true for steam keys sold on other platforms afaik

Homeowners aren’t signing contracts where they agree to use exactly 450MW of power at a constant rate 24/7 for the entire year. The problem with “Free market” utilities is that they are reliant on private sector contracts like this to fund expansion

From a business perspective, if the grid can handle the residential load 99.9% of the year, paying these businesses to cut usage during that other 0.1% of the time is a LOT cheaper than expanding their service to add one more decimal place of uptime that sits idle for the entire year

Cloud platforms like AWS/Google/Azure do something similar, where you can rent unused servers for pennies on the dollar with the expectation they can be reprovisioned by someone else on a seconds notice

Nothing more beautiful than seeing transparent yellow-orange overlaid on top of transparent orange-yellow

Top to bottom, then left to right.

What (widely popular) race could possibly be a better metric of endurance than the marathon?

It’s not uncommon to see certain sites to only work on chromium because the dev used the filesystem APIs that don’t exist on FF

Yes but if it’s first instinct is “go left” on 1-2, it’s pretty apparent the reward function could use some tuning

It’s not necessary but there is no reason not to.

Pros:

• production and development programs are more similar
• upgrading your base image won’t affect your python packages
• you can use multi stage builds to create drastically smaller final images

Cons:

• you have to type venv/bin/python3 instead of just python3 in the run line of your dockerfile

When you teach a child what a dinosaur is, you have to do a lot more explaining than when you try and teach an adult what a dinosaur is in french - the child isn’t just learning a language for those 10 years.

COVID has long lasting effects across your entire body, it impacts your immune and cardiovascular systems for years after you get it.

Risks of heart attack, stroke, etc are increased 2.5x in the year after infection, increasing exponentially with the number of times you have been infected. In fact, the risk of pretty much ANY health problem skyrockets following a COVID infection.

So I would say that if you’re explicitly trying to use Python, a Pi is the way to go.

I will point out you can run micropython on a lot of embedded boards now. I haven’t used it, so I don’t know if it’s actually good or if it’s more like those software-gore “here is my python package for building web front ends that is somehow worse than JS” packages you always see on python boards

Not quite the same, but London calling was similarly used for tourism ads among other things

Yes, but if someone trips over the cord there is a 50% chance the wrong side comes unplugged and potentially kills them, hence why they don’t make these cords

You gotta donate to planned parenthood for every dollar spent there. It’s like buying carbon offsets, but for sandwiches. /s

This seems more like a collection of examples than an actual attempt at a definition.

At its core, AI is a program that takes a given input and returns the output that, during it’s training phase, would be expected to minimize it’s error (or maximize it’s reward).