cleandreams

I once read a shocking statistic that children of the poor in London during the Industrial Revolution were 6 inches shorter than children of the rich. What raised the standard of living was bloody, persistent labor revolt. It wasn't pretty and as your comment shows it has been mostly forgotten.

It's not just issues of tech and economics but perhaps more importantly, of power. If workers have political representation (power) than these productivity increases can feed into wage increases and hour reductions. A big if. In USA, workers are pretty powerless. That means it is most likely to result in massive waves of unemployment and political instability.

When I first moved to the city in which I live there were NO homeless. That was in 1980. Since then the working class economy here crashed and the professionals took over. Now we step over bodies in the streets.

I think it is likely that the AI wave will be this, but worse.

BTW my last job was at an AI company as a high level tech contributor.

So much about this is funny. Microsoft! ha ha ha. I bet google is rolling in the aisles. AI hallucinations! I've had a view. A great way to liven up the work day. The only distasteful part is the accuracy checks that are necessary...

With you. It's excellent fun.

sometimes I wonder what havoc would break loose if we ever had a serious conflict with China.

I totally agree. I was an SAT high tester in an unsupportive home. I was not a good student. I had actually flunked a grade in elementary school though they let me pass. BTW, I ended up as a principal level at Microsoft.

When I tested 99.9 percentile in Verbal and 97 percentile in Math my father's only comment was, "Who ever thought you could test like that?" Snicker, snicker.

I'm disgusted by this decision. You HAVE to let kids have a way of knowing they have capabilities when for some reason the environment blocks that.

Microsoft bought Bonsai about 4 years ago and now has laid off the entire team. I wonder if this is related to the move to LLMs (and the new Bing). The layoffs were concentrated in AI and industrial automation.

I read recently that the fact the US military didn't want to stoke public interest with 'UFO's' meant that lots of things that were unidentified went unreported and ignored. This actually became a security risk that other countries could take advantage of. That may have happened here.

I have made a lot more than this probably 4x TOC of this and I would never share it with friends because to anyone not in tech my salary is shocking and troublesome. I have friends who are in fields that rarely top 100K and in the Bay Area it can be a challenge to live on that. It makes them sad! I didn't go into tech to make this much money but of course I am happy I've been able to take care of my family, own a house, educate my kid without debt, etc. The young ones around here all gravitate to tech in some capacity. It's hard to have a decent life with making a good income.

I am old enough to remember the hippies and related counterculture(s). (I was very young.) This is something to keep in mind: the counterculture then was a place where people made a living. And the cost of living was cheap. You could raise kids on an organic farm, sell in the local alternative stores, have a live show on pirate radio. And pay your modest bills. Health insurance and college tuition was cheap. Jeez -- in California even the UC system was free! The hippies were the pot growers back then and all of Humboldt was one big counter culture. There were lots of niche markets and vendors who could sell to them and pay their rent.

A big part of what has happened is a huge rise in basic bills (rent, health insurance, tuition) and that has forced the counterculture people to make more money. It's an economic transformation and it has narrowed our culture in many ways. Smell the desperation.

Part of that, in turn, is the slow death of the American working class. This is a much harsher society economically. That is why you step over bodies in San Francisco's Tenderloin. It wasn't like that when I moved here (early 80's).

I had a horrible manager once. He actively sabotaged me. I won't go into details. I knew about him -- 3 people left because of him and they all confided in me!! But I somehow thought I was immune. I went through a year of depression and utter confusion. I was innocent in that I didn't "get it". I had good experiences prior to him and his level of sabotage was just not something I could comprehend. Eventually I worked for a good enough company that he was demoted to a non-managerial position. But it was too late for me. I just left (I was at principal level).

A friend told me about Internet Technology Addicts Anonymous. https://internetaddictsanonymous.org/ It's quite a new fellowship and the meetings are almost all virtual / zoom. I went to a few meetings and I liked them. Right now I am trying to control my internet use with freedom (the app). Looks promising. I want to get it down to 1.5 hours/day. I have other things to do!

The internet is addictive by design and like other addictive substances (with which I have experience) it changes the brain. So I am quite serious about tackling this, one way or another.

Good luck everyone!

I feel happier reading a book. I also feel relieved when I open a book. I've been wondering why this is and I have a theory.

The information on the internet is unbounded. You know this, you are aware even if subliminally, no matter how you control your browser by blocking ads, pop ups etc. An unbounded information horizon is destabilizing. There is always more (possibly better) information a click away. It's hard to justify keeping attention on just one article. I even read a paper newspaper article with more attention than the same article online.

Of course pop ups, ads, the incredible commodification of attention, has made this much worse, but the problem is inherent to the internet.

When I open a book, there are two covers, marking beginning and end. It's awesome. In between is a place for deep attention.

There have been books that have enriched my life. It's deep focus, not shallow scanning. I think there is something soul crushing about the shallow attention we are left with after years on the job. (Slack & Teams are part of the problem.)

- The flaws in AI become more evident even as models become more powerful. Basically AI algorithms learn patterns, not abstraction, limiting reasoning in fundamental ways.

- Understanding why AI makes mistakes continues to be a black box.

- AI, despite fears, will enliven culture through DALL-e etc .

- We will be in a golden age of science and tech, not so much in fundamental discoveries but in successfully moving away from carbon based energy. Incremental improvements across a vast array of problems with create powerful synergies.

- The problems with globalization are with us for the foreseeable future, leading to increased trading among allies, less with non-allies.

- Labor unrest will increase everywhere.

- 60% Ukraine wins, 35% stalemate or increased fighting, 5% nuclear use.

- If the Ukraine situation stabilizes the stock market will take off.

It turns out there is a world wide trend towards grey, white, black, and colors in products of all kinds are decreasing. Some say it is the averaging influence of AI data analysis of consumer prefs. An overview here: https://craft-theory.com/blogs/news/are-colors-disappearing

This is sad. I find probablistic programming languages very interesting. This probably means that many of the most useful ideas are disappeared. Does anyone know any relevant papers that describe what they did?

I wrote an extremely hard and complex program in which I had to use c++ metaprogramming throughout. Not my decision. I do not understand why you could possibly need more metaprogramming than C++ provides. This is a problematic feature to begin with. Not worthless but not worth much. The debugging was difficult. Many problems show up buried in literally thousands of compiler error messages.

I don't have anything to add except this is an awesome thread. One of the best I've read.

What is the low hanging fruit in terms of changing the product? Can there be different 'streams' of twitter users? Moderation ala reddit? How will light touch moderation actually work given the garbage that will rise to the surface (child porn, beheading videos) that will kill advertising?

Yes, yes, yes. Great. Thanks. Bookmarked!

There is an interesting thread about effects on the ground in China: New rules around "US persons" are driving an "industry-wide decapitation."

https://twitter.com/jordanschnyc/status/1580889341265469440

I have written translators and they are really hard. I think training on code bases is the way to go but I am skeptical overall that it is even worth the effort. The problem is that the translation doesn't get you all the way there. The code produced is buggy and wrong in most cases. You probably get a better result from re-designing and re-implementing.

I wonder what the data being trained on even is. It has to include successful translations, right? There are not many of those. Not enough for a good result.

Admirable.

I think Stable Diffusion is a both a threat and a tool for illustrators but will mostly be a tool for artists.

The reason is that AI intelligence is probability generated mimicry. By extension it mimics art but is not art. Think of the character arcs in Shakespeare, the narratives, the layers of meanings. This sort of thing cannot even be modeled as a problem in AI. The ability to define what is necessary to produce such layers of meaning does not exist.

I think AI is missing something fundamental about intelligence. What it is missing is the ability to construct systems of abstraction and have them relate via meaning (which is informed by history, context, experience).

Something new must emerge for AI to move past probability generated mimicry.

Wow, so much negativity.

I think this is great advice. The key is being genuine and appreciative. Take care of your neediness in other ways. If you don't know how to listen, writing lots of notes can be just another way of monopolizing the conversation. Use the note as a way to maintain the intimacy you already have.

Ps my marriage lasted happily for 37 years until the untimely death of my beloved wife.

Just to note the human cost, my father died of Alzheimers. It's not impossible this fraud prevented lives from being restored and even saved.

The claim that science should be easily reproducible is wrong. Take collider experiments, or huge data set analysis problems.

I retired a bit early and I have found it immensely rewarding to take classes in various hard technical subjects. Some of these are directly applicable to jobs I could do, or have done. I think tech employers should be more supportive of life long learning. I get dinged by recruiters constantly because my resume is great but I'd rather do this, TBH. I have worked in FAANG btw. Retired at principal level.

This article seemed overly attached to bio neural nets using backprop. But backprop requires external signals that don't exist in the brain. There are other issues. Learning in the brain is structured in way that allows it to learn with orders of magnitude less data.

I think it's likely (actually I'm just agreeing with Yann LeCun) that there are basics to brain learning that we don't understand at all yet.

I liked it also.