I worked in a startup around 2019-2020 where we started using Sentry for distributed tracing. It was a company in the IOT space and we had a lot of events going through. The bill started going in the $3-5000 range just from the tracing, so we decided to host it on our own. When looking at the sheer complexity of the project, I was flabbergasted. We need a name for managing the infrastructure of company based open source, self-hosted solutions. Often times, the right choice would be to choose a different, simpler, open source solution. In this case there are some neat distributed tracing solutions, that are easier to manage.
Very inspiring comment. I suppose that software development was a much more impactful job back in the days.
I do get pleasure when building software, but like many others I also dream about starting a farm to diversify my income and get some physical work regularly.
Two nice things about F# are that you can introduce it into an organisation using the dotnet ecosystem and that you can use all libraries in dotnet, which is a huge advantage over OCaml.
Otherwise, I am happy with OCaml, but F# has also a place in this world.
You can use adapters via Foreign Function Interface and interact with C++ code. The deal breaker is that memory is separated, C++ code has its own heap and Ocaml too. Quiet different to F# in which operating with C# is seamless and the runtime is the same.
Indeed, this is also my experience. I have tried a lot of things and where quality is more important than quantity, I doubt there are many tools that can come close to Llamaparse.
What are the tools from the yesterday's world you are referring to? I've had issues with the base Python library in PDF parsing, only some state of the art tools were able to parse the information correctly.
They do not test Llamaparse on the accuracy benchmark. In my personal experience Llamaparse was one of the rare tools that always got the right information. Also, the accuracy is only based on tables and we had issues with irregular text structures as well. It is also worth noting that when using an LLM, a non-deterministic tool to do something deterministic is a bit risky and you need to write, modify and maintain a prompt.
In this vein I would also recommend: "Making the Modern World: Materials and Dematerialization" (2014) by Vackav Smil. In this work, Smil explores how these materials have shaped industrial civilization and examines trends in material use, efficiency, and potential future challenges.
Vaclav Smil discusses the main materials used in building the modern world, such as oil, steel, cement, plastics, and ammonia
I have yet to read Making the Modern World: Materials and Dematerialization however I paired his "How the world really works" with Material World. That was an excellent read in itself.
I have been a happy Ubuntu user for years. For ke the OS is completely stable and never crashes. It's also very efficient with resources, compared to Windows.
I have a music production hobby and the main obstacle for me in using Linux is no drivers for hardware and no VSTs.
reply