I’m no ML expert so take what I say with a grain of salt.
Two resources that might be useful are AWS’ SageMaker documentation and the Machine Learning Engineering book by Andriy Burkov. This book doesn’t really go into detail on logging though. One way to evaluate a model is to run a SageMaker processing job that saves the performance metrics in a json file in S3 somewhere. More info on processing jobs: https://docs.aws.amazon.com/sagemaker/latest/dg/processing-j... . AWS has various services for logging which you can look into. This will mostly apply to orgs using AWS, but it might give a sense of how things can be done more generally.
I’ve been trying to contribute to a Linux kernel subsystem recently. I can program in C, but that’s not really the barrier. Most patches are pretty tiny changes to the C code, I think. At the moment I’m finding that I don’t understand enough about the subsystem and hardware - that is the main obstacle to making meaningful changes. I’m not sure if there will eventually be AI agents that work on the Linux kernel without close supervision, but the “craft of coding” was probably never the most important thing.
Something I heard on a podcast recently re. AI is: do you think computer science, machine learning, embedded systems etc. will become more important to society in a world of advanced AI or less? I think it will become more important, but I suppose that doesn’t necessarily lead to more jobs.
I disagree. The majority of HN readers are Americans and a vocal and active subset don’t like seeing articles which are critical of their “great” nation.
I respect your opinion, but as a non-American myself (Canadian), I also really don't like having my feed filled with whatever the last drama was that Donald stirred up. HN for me is an escape from other news outlets.
There are no concrete asks from the US. The president appears to vehemently believe that tariffs will “generate revenue”. There is also the unhinged proposition to absorb Canada as the 51st state.
Almost everything you have said is wrong. Its quite remarkable to hear accusations of being ill informed or living in an information bubble from someone who is clearly suffering from both.
I hope you’re right. But even if that were the case, and trade agreements were better on paper, the net effect on the U.S. after all is said and done (damaged relationships, seeking out new markets etc.) is probably negative
Never thought that McCain or Romney would seem as amazing alternatives compared to where we ended up either. Russian propaganda brainrot really did a number on the American populace over the last decade.
You did not factor in the public opinion here. Already in Europe a mass wave of boycotting American products is shaping up. The same in Canada. It's different this time.
“The top five purchasers of U.S. goods exports in 2022 were: Canada ($356.5 billion), Mexico ($324.3 billion), China ($150.4 billion), Japan ($80.2 billion), and the United Kingdom ($76.2 billion). U.S. goods exports to the European Union 27 were $350.8 billion.”
I’m no deal maker, but it could maybe help if the US avoided pissing off its largest customer - Canada:
“ The top five purchasers of U.S. goods exports in 2022 were: Canada ($356.5 billion), Mexico ($324.3 billion), China ($150.4 billion), Japan ($80.2 billion), and the United Kingdom ($76.2 billion). U.S. goods exports to the European Union 27 were $350.8 billion.”
Depends on how you count. There are tariffs on everyone for steel and aluminium. There are specific tarrifs on Canada, Mexico, EU, and China. Trump promises to back off from NATO, which pisses off pretty much everyone on the European continent (EU or otherwise). The recent counch incident in the Oval Office wasn’t taken well in the Europe as well.
Well, there are talks about a mineral deal with Russia, others also mention gas and oil. Who knows, maybe it can grow into an all encompassing trade deal. Very good for America. The best.
My father passed away in October and it’s still difficult. I’ve dealt with grief before - two very close family members passed. The hard truth is you will never fully recover. 7 years later, a memory will bring you to tears.
The only thing that relieves it somewhat is time, in my experience. Something else that I kind of think helped was trying to guess what my dad would want for me (living sensibly, taking care of myself etc.), and try my best to follow it (not perfectly or even very well, but I try). I hope you will recover from your grief as rapidly as possible. I don’t think my father would like the idea of me being crushed by grief, and I don’t suppose yours would either!
Two resources that might be useful are AWS’ SageMaker documentation and the Machine Learning Engineering book by Andriy Burkov. This book doesn’t really go into detail on logging though. One way to evaluate a model is to run a SageMaker processing job that saves the performance metrics in a json file in S3 somewhere. More info on processing jobs: https://docs.aws.amazon.com/sagemaker/latest/dg/processing-j... . AWS has various services for logging which you can look into. This will mostly apply to orgs using AWS, but it might give a sense of how things can be done more generally.