- Can someone do an ELI5, and why this is important?
- No (human) developer would _add_ tests. ^/s
- > If a model draws a really good picture of a pelican riding a bicycle there's a solid chance it will be great at all sorts of other things.
Why?
If I hired a worker that was really good at drawing pelicans riding a bike, it wouldn't tell me anything about his/her other qualities?!
- > It's not nessessarily the best benchmark, it's a popular one, probably because it's funny.
> Yes it's like the wine glass thing.
No, it's not!
That's part of my point; the wine glass scenario is a _realistic_ scenario. The pelican riding a bike is not. It's a _huge_ difference. Why should we measure intelligence (...) in regards to something that is realistic and something that is unrealistic?
I just don't get it.
- > We are getting to the point that its not unreasonable to think that "Generate an SVG of a pelican riding a bicycle" could be included in some training data.
I may be stupid, but _why_ is this prompt used as a benchmark? I mean, pelicans _can't_ ride a bicycle, so why is it important for "AI" to show that they can (at least visually)?
The "wine glass problem"[0] - and probably others - seems to me to be a lot more relevant...?
[0] https://medium.com/@joe.richardson.iii/the-curious-case-of-t...
- > after 2 days both the cable internet and cell towers went down, so even 5G would not have helped.
I discovered the same thing the hard way myself recently (in Norway); turns out that cell towers only has enough battery for ~24-36 hours (if you're lucky).
However, someone messing with the fibre to my house is a bigger possibility than power outage, so I'll probably end up with this 5G product. :)
- If a normal person did this, the SEC would almost certainly call it insider trading.
- > humans generally have a crash, whether they are at fault or not, every 700,000 miles. Tesla has 7 in probably ~300,000 miles
This is the important part.
- I'm a "big tagger", and I have no problem with it. I don't control the type of traffic going through it, though; from my understanding, TCP/UDP wouldn't make much of a difference?
- Tide will receive Scania buses next summer (according to plan); https://kommunikasjon.ntb.no/pressemelding/18623524/elektris...
- > Hello, Clarice...
Only thing that goes on my head, but I like the car. :)
- > That way they’d finally get the status as data structures and not a hack to layout content on the web.
I remember doing this 25 years ago, but I assume (and hope) it's a huge minority who do this today?
- lol. So much coffee found the way through my nose... :D
- I haven't heard of Chroma before (sorry about that), but how does it compare to Weaviate?
- Remember that Trump doesn't like losers, so because of that "logic" Lovell shouldn't be mourned or remembered.
- > He was literally closer to God and the Heavens than anyone else before or since.
How so?
- Let's say any country create the most powerful - and thus best - LLMs. They over time infiltrate it with their political will. Over 20-30 years, I'd imagine people asking those LLMs will have their minds' shifted.
But. That's just me, my pessimism-sci-fi scenario.
- From the article:
> "Even when summing up emissions from tires, brakes, and road wear, BEVs produce 38% less particulate pollution than gas-powered cars before even considering their lack of tailpipe emissions."
- This is just bonkers; such amazing imagery! Kudos to everyone involved in this project!
> Since when are they case sensitive?
[...]