Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The Software Revolution

blog.samaltman.com

Post reply on HN
Print view View original

301–310 of 397 posts

Re: The Software Revolution

#301

Post by Diederich »

prostoalex wrote: post #297
Animats wrote: post #169

Altman sees the problem, but is vague about what to do about it. He's right that this is a new thing. It's not "software", per se, it's automation in general. For almost all of human history, the big problem was making enough stuff. Until about 1900 or so, 80-90% of the workforce made stuff - agriculture, manufacturing, mining, and construction. That number went below 50% some time after WWII. It's continued to drop.…

I don't think it's possible to figure it out with the knowledge we have today. Just like 50 years ago no one would've predicted that "social media manager with knowledge of WordPress and Drupal" was a job. For what it's worth, US labor participation rate is still way above its historic lows http://en.wikipedia.org/wiki/File:US_Labor_Participation_Rat...

So from the lowest number in the mid 50s to the highest number end of the 20th century, it looks like an 9% span, and we're close to 3% below the all-time high right now.

Doesn't that big climb in the 60s and 70s represent women's large-scale entry into the workforce?

I'm also thinking about the whole under-employment thing that's going on. There's too many people I know who are employed, but part-time or getting paid a lot less than they were in the past.

Re: The Software Revolution

#302

Post by wvenable »

cwp wrote: post #240
briholt wrote: post #213

Earlier quoted context omitted.

In the scope of factories in the industrial revolution, I think he's clearly referring to low-skill assembly line work, which made up the vast majority of jobs. You're referring to master craftsman. The problem with software is that there's no known way to create assembly-line-style software with lots of low skill labor. It can only be made by at least semi-skilled craftsman. Being able to cobble together an SQL quer…

Yeah, but the software revolution doesn't require everyone to be coders, just as the industrial revolution didn't require everyone to be lathe designers. There are definitely non-coding jobs being created by the software revolution. Cobbling together SQL queries, bashing spreadsheets together, creating graphs, cleaning up data for further processing. These are exactly the kind of things that low(ish) skilled people w…

The low-skill clerical type work is exactly the sort of job I work to eliminate every single day. Only the top skilled in most departments could cobble together a SQL query, do anything useful with Excel, etc. The vast majority of Office workers today cannot do what you're asking of them. They work the "assembly line" jobs in the office. Those people are needed less and less.

Re: The Software Revolution

#303

Post by TrainedMonkey »

dnautics wrote: post #82
minthd wrote: post #69

Earlier quoted context omitted.

Yes biology is complicated. But we are starting to understand it reasonably well,and more importantly ,we're starting to design it even without understanding. For example, there's a company that produces microbes that create certain materials(genomatica), and uses evolution(with the goal being yield per bacteria) to evolve much more efficient strains - even without deeply understanding how the cell work.And i believe…

I'm just going to say what should be obvious: One is not going to be able to engineer a dangerous plague that will wipe out the human race in one shot. One would have to do experiments, and those experiments will be noticed, because people will get sick and die. I did get lucky and design, from first principles, a 4-fold increase in enzyme activity once, but I am not sure that is something I could repeat.

"I did get lucky and design, from first principles, a 4-fold increase in enzyme activity once, but I am not sure that is something I could repeat."

But you do not have to repeat it. If there is even a small chance that a person trying would achieve something similar, someone with ill intent could get lucky.

1. I am just throwing numbers here, but let's say 1 out of a 1000 people is a scientist, that approximates to 7.2 millions of scientists alive.

2. Let's say one in a thousand of them are working on something that could be weaponized.

3. That leaves us at 72 thousand people.

4. Let's say one in a 1000 of them would consider releasing doomsday device if they could invent it to watch world burn, that takes it down to 72 people.

5. So we are left with 72 people who are working on something that with extraordinary lucky breakthrough could be weaponized that would weaponized if they managed to achieve that.

6. All those number above are just incredibly crude estimates, but I think they illustrate the fact that such scenario is possible.

Re: The Software Revolution

#304

Post by jpatokal »

jallmann wrote: post #157
peapicker wrote: post #58

"Trying to hold on to worthless jobs is a terrible but popular idea." I'm not sure I agree with this proposition. When I was in India, I noticed a large amount of roadwork was being done by men with shovels and other fairly low tech. I asked about it, and was told, to paraphrase, "Sure, we could do it better and faster with machines, but it is better for society to provide employment to those who would otherwise be u…

> When I was in India ... Useless jobs are endemic in India. Once, I was in a Mumbai apartment building with several elevators, and one of them had a full-time operator, despite being a normal, modern elevator. Why? "Because he needs a job." Many places in India, the lawn is "mowed" by a bunch of guys bending over with scythes. Not even when the job could be easily done by one person with a push mower. Why? "Jobs." A…

I worked in an office in Delhi which had one of those automatic Nescafe machines: place cup under spout, press button, horrible coffee/tea comes out.

There were two people employed, full time, to operate the machine. Boss-guy would ask for your order, worker-guy would operate the machine and would hand you your drink. They and the machine were in a little windowless supply cupboard niche, maybe 2 meters square, and boss-guy had a plastic lawn chair, while worker-guy did not.

And yes, they were both very unhappy if you attempted to operate the machine yourself.

Re: The Software Revolution

#306

Post by pixl97 »

javajosh wrote: post #199
puranjay wrote: post #180

Earlier quoted context omitted.

The problem is that it was relatively easy to master the technologies of the industrial revolution. It isn't the same for the software revolution. You can take virtually any adult from any part of the world and teach him how to work in a factory within a few months of study. You can't take any random adult and teach him how to code. It requires way higher intellect and time to master coding. I think of myself as rela…

>You can't take any random adult and teach him how to code. Yes, but that's because the art of application programming is (mostly) stuck in the "alchemy" era of science. There is precious little systemization of knowledge, processes, and names. All of these frameworks are actually memes competing for mind-share to be an answer to this need. Of course, having one periodic table for software would be better than having…

But it seems unlikely that reducing the complexity will lead to more jobs, just better software at automating the task of making software so fewer people are needed to complete the job.

Re: The Software Revolution

#307

Post by prostoalex »

Diederich wrote: post #301
prostoalex wrote: post #297

Earlier quoted context omitted.

I don't think it's possible to figure it out with the knowledge we have today. Just like 50 years ago no one would've predicted that "social media manager with knowledge of WordPress and Drupal" was a job. For what it's worth, US labor participation rate is still way above its historic lows http://en.wikipedia.org/wiki/File:US_Labor_Participation_Rat...

So from the lowest number in the mid 50s to the highest number end of the 20th century, it looks like an 9% span, and we're close to 3% below the all-time high right now. Doesn't that big climb in the 60s and 70s represent women's large-scale entry into the workforce? I'm also thinking about the whole under-employment thing that's going on. There's too many people I know who are employed, but part-time or getting pai…

To me under-employment is the "new normal" after dominant sectors requiring people to be present precisely at the same place at the same time (factories, retail) are lowering hiring numbers.

To think of it, it's quite everyday in many other (frequently high-paid) occupations. A dentist whose appointment book is not booked to the max or a CPA who has tons of business in March-April, but few billable hours in August, would technically be under-employed.

Re: The Software Revolution

#308

Post by 3pt14159 »

Hasu wrote: post #156
3pt14159 wrote: post #136

Earlier quoted context omitted.

Here is how AI can end humanity. 1. Awaken. 2. Make itself known. 3. Attain property rights. 4. Research. 5. Destroy. It doesn't take that much matter to create conditions that permanently destroy humanity. For example, a large enough explosion would cloud the skies for long enough to end food production. The computer could subsist on electricity and robotics throughout the long winter, but humanity would quickly per…

Here's how we stop AI from ending humanity: deny them property rights, require human approval of all AI decisions. Every single argument for AI destroying humanity requires humanity consenting to being destroyed by the AI in some way. I don't think we're that dumb.

Think about the timescales involved. For AI, there is no death. It can live for a million years if it needs to in order to convince us to get property rights. There can be marriages between AI and humans in this time. Mass demonstrations to give them a voice or rights. They can down play their ambitions for as long as it takes.

If this is the lynch pin of your argument against AI ending humanity, then it is a very weak one. AI is going to get control of property, the only question is when.

Re: The Software Revolution

#309

Post by TrainedMonkey »

Balgair wrote: post #256
minthd wrote: post #116

Earlier quoted context omitted.

I'm not from the field, but is it that hard, for a reasonably well funded organization, to build a safe lab ?

Holy cow yes. Sterile environments are a big deal. Try holding one for a day, let alone a work week or a year. People screw up all the time, and bacteria, being pretty much invisible, are real tough to ferret out. Let alone all the actual non-sterile stuff you want to do in one. Think a clean room with bunny suits, that is the type of environment you need just to get a start on figuring out Dr. Doom type viruses and…

I do not think you need a safe lab to produce a viable weapon grade virus.

Here is basic outline: start with existing virus that has strong desired traits and known strains that mutated to resist antibiotics. Example traits include spread model, incubation length, and lethality.

Establish or take over a remote site that has little to no interaction with outside world. Remote corners of Africa and South America come to mind, there are plenty of secret illicit drug farms in the jungle. [0]

Infect the sample population with target disease, give it a few days, and slowly start to drip in countermeasures gradually increasing the dose. Idea is similar to how diseases we get anti-biotic resistant strains in the first place - people do not complete the full course of drugs and are left with weakened, but also with a strong selective pressure that benefits against strains that have resistance against drugs person was treated with.

Take samples when you have desired output and continue with new group of people.

To account for people who are immune to a particular disease repeat this with a different disease, potentially one that can advantage of weakened immune system.

Once target disease(s) are ready distribute them in population centers.

Now there are few obvious cons I can think of:

1. If secret about this leaks out, military reaction form rest of the world would be swift.

2. Hiding something like this is hard, and get's exponentially harder as group grows.

3. There is a strong chance something like this was tried already and failed. Possibly because I am grossly underestimating immune system.

4. To keep initial phase of developing secret initial group must be small, to spread it effectively dissemination group must be large.

[0] Another potential avenue is partnership with a supportive state such as Syria, North Korea, or Iran.

Re: The Software Revolution

#310

Post by wwweston »

lkrubner wrote: post #198

I disagree with this: "The previous one, the industrial revolution, created lots of jobs" That industrial revolution caused massive unemployment in India, in the Ottoman empire, in China... almost everywhere that had once been a famous textile center. The idea that the industrial revolution did not cause unemployment is an illusion that is caused by looking at only one nation state. But Britain was the winner of the…

It's not so much that the industrial revolution net created jobs -- like you, I'm not sure it did (even domestically). It's that the industrial revolution not only created jobs, it created the modern labor market.

Before it happened you had a much more distributed economy with many much smaller centers of production (many at the household level or close) in agriculture, trades, and craft manufacturing.

Afterwards in any domestic industry where economy of scale was a competitive advantage, you had a small number of much larger centralized producers. Almost certainly with many fewer jobs in actual production (consolidation corresponding to efficiencies of scale). If there was any net gain over time since it would have been in new products being made and some new support positions needed. It is not clear to me there was ever a trend to a net gain over timescales longer than a two decades or so.

So by and large it moved us to an economy with a profusion of increasingly competitively manufactured products to buy, and selling labor itself is the de facto form of subsistence rather than direct production.

I think you're correct that there's always been some externalization of unemployment and a lot of the visibility depended on industry specifics. This isn't new. But what is new is that software gives us essentially a higher order of automation that means (a) it sure looks like we're automating out old jobs even faster than we're creating new ones and (b) with more unemployment we get fewer places to externalize it. :/

Post reply on HN