2,025 karma · joined June 3, 2016
Many discuss AI without considering that unless the LLM is going to take over the entire process, those interacting with it, must be sufficiently skilled to do the integration and management themselves.
This goes for organizations and industries as well. Which is why many companies struggle with merely digitalizing their products.
Just consider their Large Hamster Collider! It’s a travesty, and last I head, they were colliding billions of hamster per second. You also got to ask what they do with the resulting elephants and whatnot coming out of those relativistic collisions.
Stoicism promote exactly this virtue of understanding that you are in control of interpreting your own feelings.
1. Learning CUDA - the framework, libraries and high-layer wrappers. This is something that changes with times and trends.
2. Learning high-performance computing approaches. While a GPU and the Nvlink interfaces are Nvidia specific, working in a massively-parallel distributed computing environment is a general branch of knowledge that is translatable across HPC architectures.
3. Application specifics. If your thing is Transformers, you may just as well start from Torch, Tensorflow, etc. and rely on the current high-level abstractions, to inspire your learning down to the fundamentals.
I’m no longer active in any of the above, so I can’t be more specific, but if you want to master CUDA, I would say learning how massive-parallel programming works, is the foundation that may translate into transferable skills.
Let us buy our software, and separately, offer us a service agreement that actually has to provide value in its own right. The bundling of these in SaaS is what makes this obscene.
The former seems to be winning.
Sentinel 1, if I recall, is C band. But the technique would work for X-band as well, like TerraSAR-X and other commercial satellites.
The idea that it is a single-trick experiment is a misconception. Its a quantum statistical experiment where proton collisions at high energy yields random particles based on the probability of specific interactions taking place.