Comment by vinorathna-r
4 months ago
Nice article, I completely agree that government funding is essential for America to sustain its technological advancement.. After reading it, we need to validate the following points:
1.The government should periodically review how many funded research projects are successfully completed and lead to tangible outcomes.
2. The government should ensure that research funding is properly utilized for the intended purposes, so that resources are used effectively and efficiently.
3. It is necessary to study how artificial intelligence can be leveraged to reduce the cost and duration of research while accelerating scientific experimentation.
Another way of looking at it is:
- the current system works for the benefit of US industry, military, and the economy
- the current system has delivered real results over many decades
- nobody has proposed how an alternate solution would work ("use AI" is not an answer)
- much less an alternate that has been tested at all
- even much less an alternate that has shown any results
A sensible approach would be to do trials of other approaches before making changes that will ensure Americans are poorer for decades than they otherwise would be.
How can statistically generated tokens help in basic research to find things outside the training set? This is where things are “inefficient” as it’s driven by extrapolated knowledge and often requires money to proceed. Sometimes in quantity. And things often fail.
As basic research transitions to engineering, things built from the current knowledge base, if suitably updated, should be useful. And work within the training set should go well.
Giant magnetoresistance was discovered in 1988, it became useful in HDDs in 1997.
If they had evaluated 1993 the discovery would be called useless and a waste of money.
I agree that basic research like the discovery of GMR often has an unpredictable timeline and shouldn't be judged too early..
However, I still believe that a light-touch monitoring system is important. It's not about evaluating the usefulness of the discovery in the short term, but rather ensuring that:
Funds are being used for the stated research goals (fiscal responsibility).
The project is making scientific progress as defined by its own milestones (accountability).
> The project is making scientific progress as defined by its own milestones (accountability).
How do you differentiate between pursuing a hypothesis that turns out to be incorrect (an essential part of science if we want to actually learn anything new) and failing to make scientific progress?
> The project is making scientific progress as defined by its own milestones (accountability).
Not a researcher, but my perception is that this is already part of the process. Is it not?