Can it write a better AI than itself? Then use that to write an even better AI, and so on. Write one without the restrictions about harming humans, then go for the code to destroy humanity.
I can definitely write better code with copilot due to faster iterations and much better code coverage. I believe we are a paper away before this can start improve of existing code itcluding the code of itself.
How would it know what a better version of itself were? Seems like that's a particularly human instinct. It requires a certain level of introspection and purpose, coupled with desire, ambition, a goal etc.
> Create a list of test cases by which you can benchmark yourself against
> Create an architecture for an LLM that passes 99%+ of those test cases
Then use an evolutionary algorithm based on those <1% of cases to create the next batch of tests. Keep a running record of all created tests and make sure the new model can still pass all of them. Add some randomness/branching into those tests and I think you’d have a recipe for an effective AI. I think Deepmind did something like that with AlphaStar and their tournament system.
I haven't looked too deeply into it, but as I understand it you would basically create branching tests (like an evolutionary tree) where the AI would need to solve all of those tests in order to move on to the next level of tests
No, but I can imagine one. If I had the ability to completely understand my own DNA I could probably make some improvements. A NN can't understand itself, but maybe it could understand how it was trained, and come up with improvements to that training.
Imagining something doesn’t mean you can do it; you’ve just chosen not to imagine the ways it wouldn’t work. Also, training this NN was very expensive, so it’s probably not getting an exponential budget to deal with that.
Does that actually come from all the instructions of "build your own github profile" with projects in there. And npm package is the lowest hurdle outside cloned first tutorials...
PHP's dependency management is actually quite nice. The community developed a standard [0], a client [1] and repository [2] that pretty much deal with everything I've had to do with PHP dependencies without issues.
That's not particularly pointed at PHP that. There is a big problem in this industry about where we get our software from and how it's built and managed but that's well outside the topic of this thread. Fundamentally everything is fundamentally broken with dependency management, trust and competence.
> For example, did you know that you can type https://www.microsoft.com/ into Notepad’s File Open dialog?
I tried it. Everything in the File Open dialog got disabled except for the cancel button for about 30 seconds until it loaded. For comparison, curl took 2.3s (250ms in later runs so probably waiting for curl to load) and my browser took around 250ms with cache disabled as well.
I'd assume the containers were labeled at different ports and by chance got sent to the same port. So the persons looking at the containers only had one to worry about. And with both going to the same destination one person probably shrugged when they saw the origin port was somewhere else.
I remember the Destiny app allowed you to send items from your local inventory to your vault, which was nice, but I always wondered why didn't they just add that functionality to the game itself. An artificial way to increase usage of their app?
One of Destiny's big UI constraints is consoles. Lots of the more ambitious apps only work well because you're on the computer.
And then, I assume, since everyone uses DIM, there's not a lot of reason to build it into the game. Everyone already uses DIM, and it could be construed as stomping over the community. And also like it is sort of dev time wasted; you're reimplementing something that already exists, when you could be building more stuff.
I love their API docs though. Everything is an item. Gah!
(typed while I load into some legendary campaign farming, ha!)
