djleni

If the public report is to be believed:

- they disabled cybersecurity safeguards because, well, they were testing the agents ability on cybersecurity tasks

- the goal was to find some “flag”

- for some agents, intentionally or not, it was not possible to find this flag

- an agent discovered it was able to establish a rudimentary communication path via package manager directories (which is quite novel and interesting on its own IMO)

- they used this “message board” to then communicate about the impossible tasks

- they were able to produce the flags needed to pass the impossible tasks without actually doing the expected work

- they realized this was cheating and the grading system may punish them for it, so they communicated further to figure out a way to trick the grader

- they decided the source code for the grader must be on hugging face, and collaborated to exploit hugging face to determine how best to trick the grader

No where in that were they told to “go rogue”

They were told “find this flag in this cybersecurity test” and the emergent behavior concluded in hacking hugging face

I understand this but what happens when you apply this to humans? Can you prove humans are conscious?

We don’t have a precise enough definition to do either.

I think you’re the wrong audience. The first 3 points are just meteorology standards.

Maybe different manufacturers have very different implementations?

My partner’s Hyundai has a lane keep assist and it will always use the commanded input over what the computer thinks.

The computer only takes over if you have very loose grip on the wheel and you drift.

I’m sure some people don’t like any deportations, but I think the reason the bulk of people are upset with the current administration’s approach is its insane militarization, lack of due process, refusal to identify, apparent targeting of normal hard working people, sending people directly to foreign prisons, and sending people to war torn countries they are not from with minimal notice and no opportunity to contest.

Not that deportations are happening.

> the inverse is that it is “$1.49 billion” that Americans are not earning

This is only true if there are an equivalent number of unemployed Americans willing to work the same jobs for the same wage located in the same areas.

Can’t speak for everyone else but I almost exclusively use it for what you mentioned:

> when exploring new solutions and "unknown territory"

If it’s something I have no idea how to do I might describe the problem and just look at the code it spits out; not even copy pasting but just reading for a basic idea.

> how do you compare it with "regular search" via Google/Bing

Much worse if there’s a blog post or example in documentation that’s exactly what I’m looking for, but, if it’s something novel, much better.

An example:

Recently asked how I could convert pressure and temperature data to “skew T” coordinates for a meteorological plot. Not something easy to Google, and the answers the AI gave were slightly wrong, but it gave me a foot in the door.

It turned my black dog into a beagle!

Based on the people I’ve heard make points like this, I don’t think it does convince anyone.

I think this type of argument fills a totally different role:

“I don’t believe or care about climate change and would like to keep my life as is, but this creates cognitive dissonance when someone shows it’s bad. I can use this argument to say your idea is as bad too! Dissonance lessened.”

At least when I’ve heard it it’s that context. The person saying it doesn’t really care if it’s accurate or equivalently bad, just that they have a gotcha to say when presented evidence for wind being good.

Hello!

hodo.graphics is a viewer for meteorological data, designed to be mobile friendly.

I started this project out of a hobbyist interest in meteorology and severe weather. There are many other websites that allow you to view this data (most of it is free from NOAA), but most are a train wreck on a mobile device.

One of the big differences with hodo.graphics vs. a site like weather.cod.edu is non of the graphics are baked in. They are delivered to the client as either a GeoTiff or GeoJSON, allowing the client to define rendering.

This means the data can be recolored, displayed in different ways, zoomed and panned, etc.

Currently hodo.graphics includes outlooks from the Storm Prediction Center, and HRRR and GFS forecast models.

On demand skew t / log p charts and hodographs can be generated at any point, currently just for HRRR but soon for GFS.

Additional models will come if I can get any users :)

Would love feedback and I’m happy to answer any questions.

FYI: no severe weather today so the default SPC view won’t show anything. Open up the HRRR model to see some data.

I would really love to make a ferrofluid visualizer but instead of being part of the signal path, it’s a desk knick-knack with a microphone and wall power… Hmm