I don't know which it was but Dr. Jorge Diaz has an excellent video on Lagrangian mechanics as part of a series on quantum mechanics (this video just pertains to the formalism applicable classically)
I don't remember the specific video but it was pretty elementary and got across the point that I had missed, you're not looking for a global optimum through some fancy operations on function spaces, you're just doing the old fashioned calculus thing of finding a maximum by setting a derivative to zero. Except you are doing that only at one endpoint of the mystery function, and its value (the boundary value) and derivative at that point (zero) are known, and you can work out the ODE that continues the solution. That's the Euler-Lagrange equation and suddenly everything makes sense.
You can test that out. Ollama does allow you to run open source models at home. I've been playing around with a bit lately and have been really enjoying it.
On my to do list is two models running at once and building a middle layer for them to interact.
One of my fun experiments recently has been putting ChatGPT in conversation mode when I go for a walk. I recently had a 45 minute conversation where "we" fleshed out a multi-agent platform. I think a key is that you need to give each agent an "inner conversation" and criteria for when output from it gets copied to the other agents and the main chat, coupled with a process to regularly compact. I intend to set up a test system I want to run continuously, and given I enjoy working on compilers maybe I'll see how much cheaper you can do something like what OP did if you orchestrator a few agents with domain knowledge in specific areas.
I think I'd want to test a state of the art model, but it'd be fascinating to see how far you can get with Ollama as well - especially whether you can compensate for less smarts by just giving it far more runtime than I could afford with e.g. Claude.
easy to understand and use for everyone. you no longer need to spend tons of time teaching junior devs how to use git. By default mercurial also doesn't let you change history. There's no staging area or indexing before the commit. The gui and commands make total sense to the layman.
Git is definitely more powerful but causes more damage than good and wastes more time in most workplaces who are just looking for some simple version control and branching/merging features.
NetBeans was my go-to IDE until I started working with Kotlin, which more or less forced me over to IntelliJ. I’ve also tried VSCode on some pure Java/Maven projects since then, and found that it shares some of that same feel—more lightweight and direct, which I’ve always quite liked.
They're a hominid and belong to our species but are completely alien and terrify humans at a deep, genetic, evolutionary level. I love the way Watts describes Siri's involuntary reaction to the vampire, as though his fear and awareness of being viewed as little more than a potential meal are baked into his biology.
Similar to a newborn duckling that instinctively hides from shadows of a certain shape even though it has no concept of birds of prey, Siri experiences, when he interacts with the vampire, some similarly ancient, autonomic memory from the time when our ancestors were prey animals. We become little more than flighty, paranoid herd animals, jumping at the merest snap of a twig, like deer, when we find ourselves in the presence of an animal that flips the appropriate switch in our biology.
It's a wild, compelling subversion of so many sci-fi tropes and so much self-congratulatory tree-of-life bullshit and so much of our instinctual belief system regarding the way we fit into the world. It's also a completely novel (as far as I know) approach to undermining the notion of humanity's specialness, highlighting the fact that we're just animals -- and that our betters are, too, just as the invading aliens are, in a very different way.
