I detected the sarcasm but I have to say it anyways : this won’t work as long all of them are built on the same technology which is inherently vulnerable for prompt injections

Response for politics, priority shifting and upholding the morass of corporate bureaucracy. He’s worth every 100k.

• Serve the public trust
• Protect the innocent
• Uphold the law
• (CLASSIFIED)

I was a kid at the time with a pirated copy of W95. I had memorized the W95 license key (wasn’t that difficult but I finally forgot it) because I had to reinstall it so often. I think at least once a month.

Also, the LLM is just predicting it, it’s not selecting it. Additionally it’s not limited to the role of assistant, if you (mis) configure the inference engine accordingly it will happily predict user tokens or any other token (tool calls etc).

No, I guess Canada and possibly many other countries are similar but I didn’t experience them yet.

Been there twice, it was interesting and already shocking in 2008.

You US Americans should visit Europe or live there for a while. Then you can decide to come back or stay but you’ll have at least a reference point in your brain that things can be different and maybe that the US isn’t the greatest country in the world in all fields.

You always did. People walking around with guns, religious fanatics, places where you can get killed for no reason. People abducted and tortured -> Guantanamo bay. Your economy is just running better and your military is way more offensive.

I got an update for the YT Plugin yesterday evening. Also; I’m in Europe, maybe it’s a regional change?

I just tested it again, it still works here.

GrayJay still seems to work

Unlike fusion reactors AI has a pretty convincing “demo” in my opinion.

On a first glance the output of LLMs and image / video generator models is very convincing and the artifacts and mistakes appear “small” for people that don’t know much about the technical details. So it’s easy to be convinced by “we’ll just fix those little bugs and be done in half a year” promises.

EV is a similar story: electric bikes and radio controlled cars and drones work great so it’s conceivable that bigger cars and trucks would work too with a “little” battery and motor tweaking.

Nuclear fusion though isn’t really tangible yet. For laypeople or seems there is no progress at all. Every now and then some scientists report that they can hold a fusion reaction a little longer or more effective but it’s not “tangible”. That’s probably also holding back a lot of investors which with all their resources mostly still seem to invest based on a gut feeling.

As far as I know fusion energy never got that level of hype and amount of money thrown at it. I mean the research reactors are super expensive but still on another level.

I really recommend watching this introduction by Andrej Karpathy https://www.youtube.com/watch?v=7xTGNNLPyMI

One part that really stuck with me is that the data in the model is more like a fading memory but the stuff in the context window is more like the working memory. Since I learned that I tend to put as much information as possible into the context window before asking questions about it. This improved the results drastically and reduced hallucinations.

With enough RAM and ideally a good GPU you can run smaller models (~8B Parameters) locally on your own device.