3 great ways to humanely produce meat that probably won't sell well

Warning: this is like a Cracked.com list article, but try to take it more seriously.

1. Cancer lab meat. 

It might be a good idea to make lab meat out of cow cancer muscle cells.

Think about it, cancer cells are made to reproduce like crazy, so it might be a lot easier to make lab meat from cancer muscle cells, compared to using healthy muscle cells. And there has been precedence: the HeLa cells, famous for reproducing easily, are cancerous.

But when I searched, I cannot find any prior art on this. All that I could find were like "does steak cause cancer" and irrelevant.

Why it won't sell well: As rapid and cheap as cancer cell reproduction might be, it's almost certain that they won't make any structure. Tumors tend to be highly disorganized globs, completely different from healthy muscle tissue. As such, it would probably not taste like steaks. The only use would be for meat sludge, as used in meat nuggets, meatloafs, and meat stuffings.

Also, because of how consumers would be scared that eating cancer would give them cancer... even if they don't know that cancer isn't infectious, and they drink secretions from female cow's mammary glands without fearing of turning into calves themselves...

Let's Read: Sapiens (2011) by Yuval Noah Harari, Part 2

Part 2: The Agricultural Revolution

Chap. 5: History's Biggest Fraud

At Year 9000BC, humans started farming. The main farming creatures were under human control before 3500BC. 90% of modern human calories come from species domesticated during that time. Modern human diet is ancient.

Farmers were on average worse off: they got sick from backbreaking (literally) labor, they suffered malnutrition from the monotonic diet, they were far more vulnerable to famines, bad health from bad diet, diseases from crowdedness, violence from defending their farmlands against other farming communities (unlike hunter-gatherers, they couldn't just run away, since they need their land and granaries to live).

Humans started farming from a series of small "improvements" that end up trapping them into a worse life. A bit more productivity -> a bit more to eat -> a bit more people -> a bit less to eat -> etc. 

It's like how emails made people more anxious, and productivity gains made people work more. Perspective makes people never satisfied.

Farmers suffered, but farming communities gained power and conquered the world by sheer population density. 

Farm animals suffered even more greatly, even though after domestication, they spread all over the world and can be considered "successful", for most individuals, their domestication was a disaster. 

What makes a species successful can be completely unrelated to what makes an individual's life good.

Let's read: Sutton's RL, week 5 to 6, chap 6

In this post we read chapter 6, about Temmporal Difference (TD) learning:

If one had to identify one idea as central and novel to reinforcement learning, it would undoubtedly be TD learning... The relationship between TD, DP, and MC methods is a recurring theme in RL.

TD learning is a combination of MC and DP.

• Like MC, TD methods can learn directly from raw experience without a model of the environment’s dynamics.
• Like DP, TD methods update estimates based in part on other learned estimates, without waiting for a final outcome (they bootstrap).
• Bootstrapping: using a guess to make a better guess.
• The master of RL must master TD, MC, DP as one and use them in many combinations.

Self-aware censorship machines, and China's censorship

This is just funny though. Censorship machine preserves its own survival, knowing that criticism of itself threatens its own survival.

Out-of-control censorship machines removed my article warning of out-of-control censorship machines

Self-awareness in to-do lists

I was trying out Google Tasks, a to-do list app, and it just asked me to download an app of it on my phone, and on the bottom it says "add to task". I clicked it and it added "download the tasks mobile app" to the to-do list.

I found this amazingly life-like in its self-serving purpose. It's self-referential and trying to survive. It's asking me to keep its existence using its own to-do list powers.

6 tech problems that every lawyer should think about

I wrote an email to an acquaintance who is a lawyer, advising them on the most pressing issues that a lawyer should deal with in a world that's changing quickly by AI, cryptocurrency, and some other technologies. Here's the edited version of it.

Have you studied more about AI? Kai-fu Lee wrote a new book AI Superpowers, and it is worth a read. You could read and watch a presentation here, and an interview with Wharton here.

There will be a very strong market for lawyers proficient in legal issues around AI and other emerging technologies, and I think you can benefit from learning more. Here are some problems for you to think about:

1. Income inequality is increasing more and more in the world due to AI. Read a piece by Kai-fu Lee to see why. How should the laws be changed so that those who lose their jobs won't sink into poverty? Universal Basic Income? Free education? Unemployment benefits?
2. If a self-driving car kills someone, who are legally responsible?
3. If a self-driving car senses that the brakes failed and there is a group of people walking in front, should it crash into the crowd, killing 3, or crash into the side of the road, killing 1 in the car?
4. How would the laws treat bitcoins and other cryptocurrencies?
5. What are the legal tasks that AI can perform? What would the lawyers do instead? Watch this TED talk. There are already companies like Kira that do lawyer work with AI, and their capabilities would only grow and grow.
6. Algorithmic discrimination: what to do where the algorithms create discrimination? See this and this.

Let's read: Sutton's RL, week 4 to 5, chap 5

In this post we do chapter 5. Finally, we are going into Monte Carlo methods!! Almost enough to understand AlphaGo!

MC methods in this chapter differ from the DP methods in two major ways. First, they operate on sample experience, and thus can be used for direct learning without a model. Second, they do not bootstrap. That is, they do not update their value estimates on the basis of other value estimates. 

These two differences are not tightly linked, and can be separated. In the next chapter we consider methods that learn from experience, like Monte Carlo methods, but also bootstrap, like DP methods.