This is great! Really good work. It reminds me of working in a SmallTalk environment and coding inside the debugger while an exception is being thrown and restarting the computation.

I believe that this path can be supported as it is right now, and the next step would be to store a computation on some server. If an uncaught exception is raised, store all the computation along with the state, transfer it to your local machine, and restore the state of the machine as it was when the exception was thrown. This way, you can debug the state of the program with all the live variables as it was being run.

Thanks! Indeed, despite the fact that the main goal is to track ML experiments, the approach taken in `mandala` has a lot in common with e.g. time-travel debugging (https://en.wikipedia.org/wiki/Time_travel_debugging).

In reality, there are many similarities between experiment tracking, debugging, and high-level computation graphs - they're all different ways of getting a handle on what a program did when it was ran.

Probably it's because we are not training them anymore and just using with prompts. Seems like more of a swe regular type of job

except regular swe is way more fun than writing prompts

One of the best parts about Pomodoro it's that the few first session for me are about struggling with the anxiety that the work gives me. The idea behind is that the anxiety peaks and drops off rapidly in the first or second period, so just thinking that battling against the anxiety is actually work and time boxing it achieves great results

I don't know if there is a number somewhere but me and most of my peers that live in Latin America that achieve certain level of proficiency we all work for Us or Europe and not for our local economies

those are pretty good recommendations. I would start at stage 2.

Most of the time this story is true, but think this way, the person that was using the system was an expert on the subject. If you can replace the expert with a person just looking at a graph from time to time to know if you have to irrigate the soils it's a different thing. Most of the data or ML tools show us something that the client as an expert already knows, but the true power of this tools is to give them to a non expert user and have roughly the same level of proficiency

You can always get the federal state as a last term buyer if nobody wants the land.

depends, but you fine tune the whole thing, check ulmfit paper and fastai libs

What i did a couple of years algo is to approach a research lab of your interest, get in touch and try to get involved in a line of work of the lab, that usually means to work with some phd candidates, a postdoc or some junior researchers. If You stick around You should be on subject in a couple of months and ou can even get your name in some papel. You shouldn't be expecting the same involvment of a phd student, so you should keep showing up and trying to add value to the labs projects

Most of the industry people I worked with while supporting research in CS were in research labs in industry. A few collaborations were with companies who wanted something specific developed and assigned one of their group to work with us, but for more open ended research industry research labs are the way to go I think. DARPA projects often worked with us on research by pairing us with industry engineers.