krellor

I think generally the big issue that people have with crypto is that there are so many irreversible mistakes you can make, not that the underlying security is worse/better than a bank. There are lots of ways to securely manage crypto, but most people don't have the tools, expertise, and discipline to do so. Even simple things like being diligent about randomly generating strong passwords, hardening your accounts and devices against account theft and social engineering, etc.

At the end of the day if you lose your bank password, account details, etc, you can go to a branch with your id and get access. If you are scammed and money is transferred from your account, the bank will generally make you whole or be able to reverse the transaction. None of those safeguards exist in crypto, and many would say that is a feature, not a bug. Which is fine, I get it, I was a crypto early adopter because I liked the math side of it. But it's not what most people need or can integrate into their life.

I will also say that I laughed long and hard about reports of NFT smart contracts being used to execute malicious code sent as an NFT, which is a massive security issue, but I don't think it's fair to lump the whole crypto ecosystem into the NFT cesspool.

Some of my favorites:

Thinking fast and slow, Daniel Kahneman
Truly a great book that has been influential in how I approach presenting material to other people and in making sense of the world. Daniel and his long standing research partner received the Nobel prize in economics with there work in behavioral psychology. The book teaches you how people think, make decisions, and process information.

Antifragile, nassim taleb.
I won't say much other than to make a counterpoint. As much as I enjoyed the book and his presentation and arguments around making systems antifragile, his witing can be summarized by a quote from Dr. Tetlock: "His witing is like a fine French meal, gently dusted with shit." Taleb is a bit up his own ass at times, but antifragile is imo his best work.

Superforcasting, Phillip Tetlock.
Great book on how to quantify the chance of future events. Famously feuding with nassim taleb, though really it's more taleb feuding with anyone who has different ideas than him.

Man's search for meaning, Victor frankl.
One of the most interesting, heart wrenching and warming books. Whether you subscribe to his exact philosophies, frankl is a wonderful read.

The better angels of our nature, Steven pinker.
Probably the most exhaustively assembled academic book I've ever read on the trends of progress.

Origin story, David Christian.
An excellent history of everything with a focus on the repeating patterns of humanity trending towards more complex social interactions. Am easy and enjoyable read.

That's really interesting given what I mentioned in my other comment about plateauing at 4k calories/day when I was on a running challenge. Once I tracked about 4k calories burned, my energy level would just plummet.

I mean, maybe you exaggerate to make a point but there are practical limits to what you can exercise your way out of. I set a challenge to myself to jog a half marathon every day for a year. From January 2022 through January 2023 I jogged 22km a day seven days a week for 367 days. I also did light weights and exercises for my upper body. I burned around 4000/day, as best as I can track with my Garmin watch. Which throw in a couple of milkshakes and you can blow through 5k calories in a day.

I will say, I did struggle to keep my weight up with such a regimen and a fairly healthy diet and dropped to 150lbs at my lowest (6'2"tall). But if I wanted to eat more calories I could easily get there with fried food and ice cream.

Edit: and for most people this is completely infeasible. Most people don't have the time flexibility to wake up at 4 am every day and put on those kinds of miles.

I don't know where op is from but a lot of states, especially in the western US require invasive species tags on any craft over a certain length, like 10 ft. Some also require state registration decals. Idaho is an example of one that takes the invasive species decals and water craft checkpoints very seriously.

Just moved into a house with ac for the first time and it is well insulated and lots of shade from trees. At night before bed I set it to 68, and in the morning I set it to 74. Even when we had 100 degree days it never got above 73 inside, so basically I only run the AC at night.

I'm sure you mean all the heights are a little different, but I'm envisioning where every step has to be taller than the previous one and the shenanery that would happen to make long stair cases navigable. Start with teeny tiny steps to end with uncomfortably large ones.

These sort of thought experiments are helpful to drive home the point that in creating public policy, sometimes you do need a way to quantify the value of human life, which gets to the ugly truth that we do value lives differently everyday in society. ER triage will save the sickest person first, all else being equal. But when things get swamped children often get prioritized up. We value young life more. The trolley problem forces the concept of assigning value to life and taking action based on that.

This second problem doesn't do so as cleanly. In my opinion, the right answer is to let them riot, but also alert the police and seek to mitigate the harm they cause. This problem feels less about the quantification of human life than about moral culpability of actions. The trolley is acting as it must die to physics. It isn't good or evil and has no agency. On the second problem, the rioters have agency, are choosing to do evil, and should be fought against.

Just my quick take though.

They had a bug where if you were denied it was supposed to send a notification with a link allowing you to elaborate, but those notices didn't go out. Then, when you've applied but gotten stuck in the unapproved state, your user account exists but is disabled, this you can't make a new account with that same name. Sort of a sucky situation all around.

I don't think it's as much that the meta model was replicated as much as they fully open sourced it with a license for research and commercial use.

I actually think the market demand will be fairly small for fully offline AI. The largest potential customers might be government who require full offline hosting, and there is a small group of companies servicing that niche. But even government customers who require that their data is segmented are simply having enclaves setup by the big cloud platforms where they guarantee that inputed data isn't fed into the training process and doesn't leave the customer environment.

I fully support folks who sustain open source AI frameworks, but in terms of commercial customers that will drive industry trends with dollars, I expect there will be demand for hosted solutions that use proprietary models.

This is not a picture of happy neighbors, lol.

For most people computers are just the same as cars. People want a car that will drive them from place to place, are easy to refuel, easy to operate, and can be taken to an expert for anything difficult or that requires specialized knowledge. Same for computers. Most people want a computer to navigate the web, install the apps they are used to and that their friends use, is easy to operate, and can be taken to an expert for any involved work.

Even the friendliest of Linux distro don't check all those boxes. You cant get ready support from a repair shop, many of the apps are different or function differently, and it doesn't receive all the same love and attention from major third party developers as Windows does.

Most people could learn to use Linux; it's not that hard. Most people could learn to change their own oil. But for most people, it's not worth it. For most people it's not the journey, it's the destination and cars and computers are just tools to get there.