The power of FORTH is that creating a variation tailored to your specific problem is almost effortless.
In most languages, building a compiler or interpreter is a major project. With FORTH, if you’re the only user, you can have something working in minutes.
I’ve lost count of how many times I’ve thrown together a simple stack-based interpreter. Whenever I need to encode non-trivial behavior in an app, I know I can spin up a quick FORTH and get it done.
I did. And when I went open with it, I did find out that plenty of people whom I know had breakdowns too.
The hardest thing was to admit I needed help (used to be a high performer: I solve my own problems!). It was very late, and by then it took a couple years, therapy and medication to recover. My doctor was very optimistic though. These sort of breakdowns happens to be very typical, and he was convinced that I would do a full recovery. I was skeptical. He was right.
The second hardest thing for me was to understand that it wasn't my fault. Athletes break their muscles often due to overwork. Brains can also break. In both cases recovery requires time, and often some treatment.
>So HN, what has your experience been with depression?
I am sharing my own experience, the topic being a taboo does not help people who suffer.
I was also shit scared of medication, and too apprehensive to go to a medical professional... until it was too late and it all collapsed.
I visited 4 different emergency psychiatrists. Only one of them managed to identify the condition and treat me accordingly. His goal: keeping me alive for 18 months. It worked. I did need strong medication, and it had to be carefully adjusted several times. This kicked me out of depression in about six to eight months, but did not solve underlying issues. Once the emergency psychiatrist rated that I was in a stable route for recovery, I switched to a regular psychiatrist that basically checks my medication once every two months.
In parallel, I started weekly therapy with a psychologist. Also I tried two doctors and ended up with the second one. This one is needed to solve underlying issues. It took long time to get it really working, but it ended up great. It's been a year and I am still attending weekly therapy.
Its been 18 months since I started treatment. I am better now than I've been the last 10 years, but change did not come overnight, and I am still working on it with full energy. I wouldn't have made it so well without any off:
a) sport & good food & sleep.
b) amazing doctors.
c) medication.
> Have you tried the drugs?
Yes, quite a few actually.
> What worked or didn't?
What worked: medication + full trust in the doctor + lots of patience.
What it did not work: expecting to be back to "normal" fast.
> Have you been able to triumph without chemical assistance and what did that look like?
I really doubt I would have triumphed without chemical assistance.
> Is my utter terror of these drugs warranted or should I just bite the bullet and try them?
Terror is due to depression being taboo, and fear of the unknown. Medication is not to be underestimated. It has side effects, for me it was mostly feeling even less energy than normal. So I would not take them unless you have under close surveillance and you trust fully your doctor. On the other hand, they are very very effective. I am weaning-off medication slowly and I am experiencing no side effects.
I can relate. I just read this at 3am while trying to stop wondering what to do with my life. I had the childhood dream of being a professor, and after a short stint in industry, when someone gave me the chance to do a PhD I took it. Life got in the way and got my PhD at 36 years old. Nobody guided me, and did not optimize my publication schedule, so it is impossible to land a job in academia. Now I am postdoc 39, non hirable bc I'm old and my cv compares badly to anyone elses. No clue how to provide for my family.
Last month we helped clean the room of a colleague after he "passed away" after 10 years of post doc.
Another colleague got a job at 36 as a bare programmer after a very successful PhD where he wrote books and was invited to conferences. At 40 he rage quit this job after a bout of frustration and we haven't heard from him since.
Luckily they didn't have kids.
Wouldn't I have kids I would definitely follow their path.
I'm tying (unsuccessfully) to get sleep, and I'll try magnesium supplements too. I already spoke about this but never got solutions, and at this point it feels almost even worse every time.
If you have any resources on how to manage anxiety please bring them on. My old son has a mild development impairment, I am the sole bread earner, and we have no family support (we live abroad).
While the roots have been known for a long time, my impression is that the key paper that started this line of thought was Marco Cuturi's NIPS 2013 paper "Sinkhorn Distances", which is, IMHO, a very nice read.
IMHO the coronavirus was used as an excuse to kill a fair that was already producing negative ROI.
Due to the size and fame of the fair, most big companies were forced to participate to "look better" than their competitors, even if they didn't have any great novelty to show there: it was a marketing arms race where each company had to invest more than the previous year.
I'm not saying to discard this research. I'm suggesting to wait until it is peer-reviewed and published before jumping on it.
To me, the capsule concept seems reasonable, and I have my personal opinion about its strengths and flaws. But my opinion hardly matters.
I expect peer reviewers from NIPS to have a better understanding that I have, and I trust them to filter and clean this idea, instead of trusting the research just because of the name that signs the paper.
To me, although it has its flaws, the _double-blind_ _peer-reviewed_ processes is important.
Nice video, however instead of riding the hype train of arxiv, could we wait until peer review analyzes the paper?
If someone other than Hinton presented a YADLA (Yet Another Deep Learning Architecture) that does not achieve state of the art level of performance in the basic datasets, it would not be very well received.
I agree, it still needs a strong use case. IMHO, new network architectures are generally overhyped. Even AlexNet performs wonderfully well in most problems once you add nice initialization and batch normalization.
I am very disappointed with constexpr in C++14 and C++17.
The work shown here is impressive, but the point is that it should not be. D has shown that variable initialization at compile time can be painless.
The main limitations I find are:
a) Lack of support in the standard library (e.g., you should code your own sort, vector class, etc.)
b) Terrible Terrible Terrible compilation times.
c) Lack of a dynamic memory allocation within constexpr.
constexpr functions are not evaluated while parsing (e.g., the code is not compiled and then executed). My use case, which was to avoid a 1 second preprocessing time while loading a library, took more than 20 minutes to compile and used more than 60 GB of RAM (on GCC7, CLANG did not even manage to compile).
Stuff like initializing a bitset can easily eat all your ram (e.g. this fails to compile):
#include <bitset>
int main() { std::bitset<102410241024> bs; }
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63728
In most languages, building a compiler or interpreter is a major project. With FORTH, if you’re the only user, you can have something working in minutes.
I’ve lost count of how many times I’ve thrown together a simple stack-based interpreter. Whenever I need to encode non-trivial behavior in an app, I know I can spin up a quick FORTH and get it done.