title: Weekly-#12 summary of LLM acceleration tags: [Weekly] comments: true
Almost finish it
start the small business
Ensuring accuracy of AI-Generated contents is always challenging, running codes to check if it's right is a common method in coding field.
This works at least in competitive programming contest, if the solution is not corrent, we can feed in with corresponding test case, which will help it to correct the answer. More important, there is a 100 percent correct test cases system for the problem, which make AI works all the time.
Taking this in the true world, All we need to are how to evaluate the work.
Anything existing must be reasonable, almost everythink has its reason.
There should be a methodology for everythink, because even building product have one.
sloth is so cute.
The world is a simplified animals world, with small differences. Some people will go to zootopia, other will stay in rabbit town.
Asking for several days, finally got one invitation. This blog definitely contribute more for it. Power of personal product.
Read more, Do more, Think less