All of Tim Freeman's Comments + Replies

Found a different, perhaps better explanation: salt intake leads to temporary weight gain from water. Restaurant food is salty. 42:30

That explains noise but at the time I had almost no variation and then sudden drops. I was also consistently high salt, and I expect the salt effect has a ceiling.

I tend to have itchy eyes. An optometrist suggested "derm dry eye relief mask" by eyeeco. Heat it in the microwave 20 seconds or so and then lay down with this lump of warm stuff on your eyes for 10 minutes until it isn't warm anymore. Gently rub once afterward.

This seems to help and I do it fairly reliably.

The theory is that there are glands on your eyelids that secrete some magic substance that makes your eyes dry out slower. Those glands get clogged up if you stare at a computer screen and don't blink enough. Rubbing at them does a poor job of removing ... (read more)

Yeah warm compresses are nice and I guess I recommend them as well, especially for relief when symptoms are severe. They are a bit of a hassle though so I recommend looking for other solutions.   The best product for compress that I have found: [] It gently steams your eyes, so it makes them both warm and moist. I think I do recommend some form of moisture when you do warm compresses.  

In response to "how so?": If this catches on, you can sell a drug by infiltrating the forum and posting fake news of miracle cures under many different names.

For schemes like this to work, you need some way of guessing who trusts whom. The spammers might claim to trust each other, and you never really know who the spammers are. The best you can hope for is for the real people to get good information from other real people they trust, and the spammers get garbage information from other spammers but that doesn't matter because they are spammers.

I don't know of any implementations of this.

I read the book before reading this review. I have recently had success with the Conference Therapy technique they describe, so I highly recommend the book.

I actually started reading the book, rage-quit in the middle, then came back to it years later and found it useful. I rage-quit because the section on EMDR was about a patient with panic attacks, EMDR was done, and afterward the patient still had panic attacks but they claimed the treatment was a success anyway. Any sensible interpretation would call this failure. So at least one of the authors does mot... (read more)

Cool! Huh, I didn't remember this from my read. Searching for "panic attacks" in my copy now, there's the story of Susan who got EMDR for panic attacks, but my copy seems to say she only had one single panic attack after the treatment and after that they stopped for good? Is that the section you mean? Can you say more about solving the problem multiple times? Is it maybe partially explainable by the same schema having been stored in the context of many different situations with each of those needing to be reconsolidated separately (something that the authors do mention), or being based on multiple different experiences that need to all be addressed before the symptom goes away entirely?  My reading is that "symptom" means something that disrupts a person's life enough that they seek therapy for it - if there wasn't anything that they experienced as a problem, they probably wouldn't come to therapy in the first place. So I interpret "symptom" to suggest that some of the beliefs are disruptive, even if not necessarily false. I don't remember whether this was in UtEB, but Coherence Therapy: Practice Manual & Training Guide explains the concept of a purposeless symptom that's the byproduct of something with more function. An example they give is a belief that a person needs to hide from the world, which limits their ability to act so much that they become depressed. In that case, the depression is a symptom, but it's based on a correct belief that the person is currently unable to do things that would get them what they want. Of course, it's true that the methodology presupposes that there's some incorrect belief in the client's system. Otherwise there wouldn't be anything for memory reconsolidation to fix, and those beliefs are somehow linked to the symptom. But I don't read that as contradicting the claim that when you start investigating, you're not going to know exactly which one of the client's learnings are true and which ones are false. Hmm that makes sense

In response to: "I don’t understand why or how weight-loss-that-is-definitely-not-changes-in-water-retention comes in chunks. If you have an answer I’m quite curious."

I too have observed that this happens. I read somewhere that if you lose fat, it is a few cells losing all of the fat instead of many cells each losing a little bit of the fat. The empty fat cells fill with water and your weight stays approximately constant. After there are enough empty fat cells that have been empty long enough, some of them do apoptosis and you pee out the water and you lose weight then.

I don't remember where I read it.

I read a similar thing on Reddit repeating something the author's trainer said once. I have almost zero confidence in this explanation and it's also the best I've found

Does the discriminator get access to the symbolic representation of f, or just its behavior?

If the discriminator only gets access to the behavior of f, I might have a solution. Define g(y,z) = 1 if y = y * z, 0 otherwise. So g(y,z) is 1 if z is 1 or y is 0, which is two different mechanisms.

Pick some way to encode two numbers into one, so we have a one to one mapping E between pairs (y,z) and numbers x. Define f1(x) = g(E(x)).

Now pick a cryptographic hash H that might as well be sha256 except some fresh algorithm not known to the discriminator or guessable... (read more)

We're imagining that the discriminator gets to see the computation done by f---the tricky part of the Fermat test is that even if you see the computation of an(modn) you still can't figure out whether the number was prime or Carmichael. We are also hoping the discriminator will be no harder to learn than f, and so e.g. if gradient descent bakes some tricky secret into f then we are happy baking the same secret secret into the discriminator (and this is why obfuscated circuits aren't a counterexample).

Eby hasn't updated his blog in 10 years. He didn't even put a post there describing his wonderful next job or project, or his wonderful retirement. He is posting to Twitter, so he's not dead.

If his advice worked for him, he wouldn't be in that situation.

That's an old blog, he's currently active on [] and []