Happy new year all!
17 Likes
16 Likes
Happy New Year everyone! May it be a good one!
13 Likes
1³+2³+3³+4³+5³+6³+7³+8³+9³. Yup, checks out.
9 Likes
Happy New Year to everyone, I hope you had a good evening and a good time yesterday!
13 Likes
Happy new year all 
10 Likes
There’s a theorem that says the sum of the cubes of the first N integers is equal to the square of the sum of the first N integers. So
2025 = (1 + 2+3+4+5+6+7+8+9)^2.
7 Likes
They were, though.
(Carbon zinc aaa cells, for use in a rarely used and absurdly expensive and required remote control.)
8 Likes
This is one of my mum’s favourite images from a decade ago, 4 generations of us all together, from my 102 year old nan to my 4 month old son.
My partner was having a look at improving the quality on Photoshop and the AI helper popped up and asked her what she wanted.
Interested, she asked it to remove the blurriness and have everyone looking at the camera.
‘No problem!’ it replied, and produced this image.
For some reason, I am thinking of a liquid that tastes almost, but not quite, entirely unlike tea.
20 Likes
Surely AI will be coming for our jobs any minute.
9 Likes
Nonsense! That’s just a hair extension.
3 Likes
AI doesn’t need our jobs; it just needs access to the nuclear weapons. It’ll probably be way better at launching them than it is at editing photos.
(Always start the day with a jolly thought; that’s my philosophy…)
8 Likes
AI already has its hands on all the “nuclear” codes for the Internet… and it has launched all the algorithms it needs to destroy the “us” by dividing (and then conquering?) us into smaller and smaller tribes and bubbles. All the people warning of future AI-pocalypse are disingenuous because they are–probably knowingly–ignoring the terrible effects we’re already seeing from what most people don’t realize is of the same type of machine learning AI that powers the latest (hype-)cycle.
AI has first come for our minds.
3 Likes
I asked an ai for a recap of 2024. It told me nasa successfully landed astronauts on the moon. Guess I was not paying attention to the news.
6 Likes
And how many MWh did that take?
5 Likes
That’s a surprisingly hard question to answer, even if you have access to the providers energy use data. To run the inferences, and generate the result, the answer is going to have 10^-5 as the order of magnitude. But training the model is vastly more expensive, and some of that should be attributed to it. That will increase the magnitude. How much depends on how the allocation is made. Wouldn’t surprise me if it moves it to 10^-3.
2 Likes
Hi. If the Great Vowel Shift is a person, I would be in jail for First Degree Murder.
6 Likes