Wow that is super vitriolic, more if from a colleague. Sure the hehe debate is about if he is to be considered more or less accomplished, but going from non-success to pariah, what is the need?
> but going from non-success to pariah, what is the need?
Not sure "need" is the appropriate word here.
Grab anyone in who has worked on AI in the last 30 years, and pronounce the word "Schmidhuber" and watch the face of you interlocutor: you'll either get an eyroll or a smirk, but rarely a lively discussion on what he's "invented".
This is understandable, but maybe you cannot offer an alternative to just give credit nobody. Schmidhuber is playing the game by its rules being self-consciously contrarian enjoying himself, while offering an interesting historic pinch of retrospective justice.
That is the price you pay. If you refuse to pay you are left to express a potentially complex state machine in terms of a flat state-transition table, so you have a huge python cases statement saying on event x do this and on event y do that. That obscures evident state-chart sequentiality, alternatives or loops (the stuff visible in the good old flow-charts) that otherwise could be mapped in their natural language constructs. But yes, is not honest flow. Is a tradeoff.
Heisembugs aren't just technical debt but project killer time bombs so one must better have a perfect thread design in head that works first attempt, else is hell on earth. I can be safe in a bubble world with whole process scope individual threads or from a thread pool (so strong guarantees of joining every created thread) and having share-nothing threads communicating only by prosumer sync-queues that bring a clear information-flow picture. One can have a message pump in one thread, as GUI apps do. That is just a particular case of the prosumer channel idea before. Avoid busy waits, wait on complex event conditions by blocking calls to select() on handler-sets or WaitForMultipleObjects(). Exceptions are per thread, but is good to have a polite mechanism to make desired ones to be potentially process-fatal, and fail earliest. This won't cover all needs but is a field-tested start.
TBH it is not ALL cloudFlare IPs but a significant quantity of sites using and not using CF CDNs. You cannot imagine what a pest that is even for legit users of legit collateral damage pages. CloudFlare is in the courts appealing/countering initial court allowance to blockade and ISPs are bound to comply to blackout requests. You can look at https://hayahora.futbol (traslation: is there soccer match now?) to see affected domains.
My 2 cents is they do justify it by the interest of the consequences, as Tychonoff or Nullstellensatz. I wouldn't call that faith: Best practices is to state Tychonoff as "AC implies Tychonoff" and that last is logically valid. Sometimes the "AC implies..." is missing, buried in the proof or used unawaredly or predates ZFC, and is a bad thing. But very ofen one now see asterisks on theorems needing it.
In a complementary story, from light reading I gather that Zellig Harris, that advised Chomsky's dissertation, contributed decisively to distributional semantics, a forerunner of word embeddings, in turn a key to the door of neural NLP. The ML guys had a shiny toolbox ready to use for decades. Had it been more fashionable...
I am testing it as a replacement of MathPix, first few tests look rather decent. In python for windows: https://pastebin.com/uyiFHKdJ (alpha version prototype). Launches windows snip tool, waits for clipboard image, calls Mistral, retrieves markdown and puts it as text in the clipboard, ready to be pasted in Typora, Obsidian, or other markdown editor.
reply