The fact that we're using AI killer robots to wipe each other out in droves doesn't bode well for that future does it...
Why do we watch Olympic runners, when cars on your average city street easily exceed Usain Bolt's top speed on their morning drive to Starbucks? Why do we watch the Tour de France, when we can watch Uber Eats drivers on their 150cc scooters easily outpace top cyclists? I'm sure within a couple years a Boston Dynamics robot will be able to out-gymnast Simone Biles or out-skate Surya Bonaly. Would anyone watch these robots in competition? I doubt it. We watch Bolt, Biles, and Bonaly compete because their performance represents a profound confluence of human effort and talent. It is a celebration of human achievement, even though that achievement objectively pales in comparison to what our machines can accomplish.
I think the same is true for other aspects of human creativity and labor. As we are able to automate more and more, we will place increasing importance on what inherently cannot be automated: celebration of our fellow humanity. Another poster wrote that "bullshit jobs" [0] exist primarily because we value human contact [1]. I am inclined to agree.
They are not "bullshit jobs"
They will become so only after the day when AI "help" and "support" is actually better than talking to a human.
Which is not happening anytime soon, possibly never. Call me when it happens
So, sure, there will be space for some human achievement for the sake of it, but, most fewer and fewer people will make a living off that.
Olympic Athletes are a combination of luck in the genetics department and a lot of effort, but ultimately do not seem to be sufficient to help the athletes themselves.
Big sports events are the "circenses" part of "panem et circenses" [1]. Fun fact concerning this: the German word for "entertainment" is "Unterhaltung"; thus it can be argued that the purpose of entertainment/Unterhaltung is "unten halten" (to keep at the bottom), i.e. to keep the mass of the populace at the bottom, or in other words: to prevent the mass of the populace from coming up.
> Would anyone watch these robots in competition?
I have seen robot fight competitions both live and in videos, and I have to admit that these are not boring to watch.
So yes, with a proper marketing I can easily imagine that lots of people would love to see broadcasts of some robot competitions.
--
No, that would be "Untenhaltung", which isn't an actual German word, but could be.
"unterhalten" in German can both mean to entertain (however, not as in "entertaining a notion") having a conversation, as well as "to maintain". It has several meanings, all of them positive.
When chess engines started becoming really good, some people worried that competitive chess would die. Today, grandmasters stand no chance against a smartphone, and yet, chess popularity is at an all time high.
(https://news.ycombinator.com/item?id=47587863) A comment I had written sometime ago. Aside from a very few at the top, I have seen some chess players regret in a very nostalgic way.
The chess industry continues to allege against each other and we lost a star (Rest in peace, Daniel Naroditsky) because of it. The current world champion himself is struggling from all the pressure put on a 19 year old boy.
We enjoy playing against each other but man it is competitive if you wish to feed families.
Most of us play chess out of leisure. I am unsure how a world where everyone does something akin to chess competitively (ie. for money, as we wish to feed our children and ourselves) would look like.
One can say something similar to UBI might be needed and then we all play chess in leisure, but I don't think that is what most people propose when they mention the example of chess.
Chess is an unusually poor example. When computers took over Chess, we didn't have something stupid like 30% of employment relying on playing Chess to eat and pay rent.
The analogy only makes sense if you're already convinced that we won't lose the majority of white-collar work to computers.
To those who are not convinced that we are looking at making 50% of the workforce redundant, Chess is an analogy that makes no sense.
It only makes sense if you're already a true believer.
All of those sports make intuitive sense to me, I really don't get why we make such a big thing of balls though.
F1 is somewhat about which company can build a better car. But any real improvements seem to invariably lead to a rule change that bans that improvement in future seasons. So you are back to drivers being the most visible differentiator
There's still space for creativity, novelty, invention and human intuition.
40 years ago, there was a market for:
* newspapers
* cameras
* navigation tools
* HiFi equipment
* photographers, translators, etc
Same thing will happen for all the other markets that AI will gradually eat. Sure, you can find a human that can do better. But that costs 90$ / hour and requires finding someone, negotiating a contract, etc. But when people can do something good enough in 30 seconds with something they already have access to, and move on with their life, then that's what they'll do.
So just raising the floor will have a big effect on society.
We haven't needed the overwhelming majority of human creativity. We still paint and play guitar even though it has no economic value. I think we'll continue to do these things regardless of AI.
> and work
This is another story.
Are the only options here being a good and "useful" worker/consumer, or a violent, irrational thug? Is there nothing else you can imagine?
People also need their lives to have value. We are social animals. As a generalization, there is a strong desire to be (viewed as/able to view themselves as) a contributor to the community.
These don’t have to be linked: we have (significantly!) stay-at-home-parents and philanthropists and retired community workers. But in our current values system, it is often linked - having a job in the household is viewed as a moral good. It might be hated, but it’s at least “contributing” something.
If this goes away, and we have millions completely adrift? With no structure to contribute to? Even with the largest welfare expansion in history, I think we’re preparing for a very turbulent society.
But what I worry about sometimes is when you snatch that away, then you just lead to stress over basic existence.
> If this goes away, and we have millions completely adrift? With no structure to contribute to? Even with the largest welfare expansion in history, I think we’re preparing for a very turbulent society.
Please look around and just try to remember how many things have happened in a year or two, We are already within a turbulent society but yes I also feel like this isn't the end and the cat is sort of out of the box and the world has to prepare itself for even more turbulences/radical changes.
This whole prescriptive thing this response and others have where its like "ah surely it is up to us to find some meaning for the masses of plebs in our brave new world" is, IMO, presumptuous at best.
Like literally just give people an actual chance to find their own meaning, and I promise you they will find it. If it seems hard to you or "full of turmoil", that suggests a poverty of inspiration on your end, not everyone elses. Meaning is not intrinsic to our particular mode of production at the moment, in fact, individuals find meaning despite this mode!
Yeah, this is not happening anytime soon. Have you even looked at AI-generated code or text? AI is just a dumb parrot, it's no match for human effort and creativity even in these "easy" domains.
The business case for AI generation is just being able to generate huge amounts of unusable slop for next to nothing. For skilled workers it's a minor advantage in that they get a sloppy first draft that they can start the real work on - it makes their work a bit more creative than it used to be, by getting rid of the most tedious stuff.
You really need to look again. If you're still manually writing code you have your head in the sand.
AI can produce better code than most devs produce. This is true for easy stuff like crud apps and even more true for harder problems that require knowledge of external domains.
It just makes you MORE of whatever it was you already were.
They're doing things now that they either flat out could not do before, or if they did it would be an giant mess (I realize they still can't really do it now, AI is doing it for them).
I'm not sure about other devs, or even their number, but AI can most definitely NOT produce better code than I can.
I use it after I have done the hard architectural work: defining complex types and interfaces, figuring out code organization, solving thorny issues. When these are done, it's now time to hand over to the agent to apply stuff everywhere following my patterns. And even there SOTA model like Opus make silly mistakes, you need to watch them carefully. Sometimes it loses track of the big picture.
I also use them to check my code and to write bash scripts. They are useful for all these.
That's fine, and useful, but you're really putting a ceiling on it's potential. Try using it for something that you aren't already an expert in. That's where most devs live.
Even expert coder antirez says "writing the code yourself is no longer sensible".
And he didn't limit his take to just C code. He said: state of the art LLMs are able to complete large subtasks or medium size projects alone, almost unassisted, given a good set of hints about what the end result should be.
These things bullshit their way about all the time. I've lost track of how many times they seem to produce something great, only for me, upon deeper inspect, to see what a subtle mess they have made. And when the work is a bit complex, I cannot verify on sight; I'd have to take time to do it.
Also, they absolutely cannot even produce some levels of code. Do you think I can just give them a prompt to produce a haskell-like language, allow them to crank for some hours, and have a language ready made?
Want an example? here is something Sonnet gave me just today:
const sort = sortKey ? { field: "name", order: "ascending" } as const : undefined
const sortKey: "name-asc" | "name-desc" | "recently-accessed" | "least-recently-accessed" | undefined
Here is another one:
-------------------------------
Given:
queryX: <Ent extends EntityNamePlural, Col extends StrKeyOf<Dto<Ent>>>(args
: {
entity: Ent
query: QueryArgs<Dto<Ent>, Col, fOperators>
auditInfo?: AuditSpec
}
) => Promise<Result<Pick<Dto<Ent>, Col>[]>>
export type QueryArgs<Rec extends StdRecord = StdRecord, Fld extends StrKeyOf<Rec> = StrKeyOf<Rec>, FltrOp extends FilterOpsAll = FilterOpsAll> = {
/** Fields to include in results (defaults to all) */
fields?: Fld[],
/** Filters to apply */
filter?: RecordFilter<Rec, FltrOp>,
/** Sorting to apply */
sort?: {
field: Fld
order: SortOrder
},
/** Pagination to apply */
page?: {
maxCount?: number | undefined
startFrom?: {
sortFieldKey: any,
idKey: ID
} | undefined
}
}
const sort = sortKey ? { field: "name", order: "ascending" } as const : undefined
const xx = storage.queryX({ entity: "cabinets", query: { filter, sort, page: page ? { startFrom: page } : undefined } })
Which is obviously wrong. I should be getting the full type, i.e., all columns picked. The problem is that the Column generic parameter is not being properly inferred, which is (probably) due to the sorting by name, since the sort column is defined to have to be part of the query field name, so when field is not provided, TypeScript infers the fields as the sort column name.
Neither ChatGPT nor Claude Opus have been able to solve this after one hour, suggesting all kinds of things that don't work. But I have solved it myself, with:
export type QueryArgs<Rec extends StdRecord = StdRecord, Fld extends StrKeyOf<Rec> = StrKeyOf<Rec>, FltrOp extends FilterOpsAll = FilterOpsAll, Srt extends Fld = Fld> = {
/** Fields to include in results (defaults to all) */
fields?: Fld[],
/** Filters to apply */
filter?: RecordFilter<Rec, FltrOp>,
/** Sorting to apply */
sort?: {
field: Srt// StrKeyOf<Rec>
order: SortOrder
},
/** Pagination to apply */
page?: {
maxCount?: number | undefined
startFrom?: { sortFieldKey: any, idKey: ID } | undefined
}
}
queryX: <Ent extends EntityNamePlural, Col extends StrKeyOf<Dto<Ent>>, Srt extends Col = Col>(args
: {
entity: Ent,
query: QueryArgs<Dto<Ent>, Col, fOperators, Srt>,
auditInfo?: AuditSpec
}
) => Promise<Result<Pick<Dto<Ent>, Col | Srt>[]>>Passed some point, if you are good at what you are doing, the AI will stop helping and become a burden, because you will want precise control, and AI in its current form (deep learning) is not good at it.
There is a reason we talk about "AI slop", you simply cannot let an AI make creative decisions and expect a good result.
By creative I don't just mean artistic. For code, AI works for the least creative tasks, like ports, generic-looking CRUD apps, etc...
As for work, we have already eliminated most of the need for human work. By "need", I mean survival: food, shelter, these kinds of thing. Most of human production goes to comfort, entertainment, luxury, etc... We will find stuff to do that isn't bloodshed. In fact, as times went on, we spend more on saving people than killing them, judging by a global increase in life expectancy. Why would AI reverse the trend?
I don't think we're anywhere near that point.
The funny thing is that I am a sort of misanthrope. And in that, in this forum, I seem to have a lot more respect and optimism for human potential and ingenuity than the majority here.
I can see two major delaying factors here:
1. Current generation LLM technology won't scale to true AGI. It's missing a number of critical things. But a lot of effort is being spent fixing those limitations. But until those limitations are overcome, humans will be needed to "manage" LLMs and work around their limitations, just like programmers do today.
2. Generalist robotics is far behind LLMs for multiple reasons, including insufficient sensors and fine motor control. This would require multiple scientific and engineering breakthroughs to fix. Investors will, presumably, spend a large chunk of the world's wealth to improve robotics to replace manual labor. But until they do, human hands will still be needed in the physical world.
The real danger is if AI passes a point where it starts contributing substantially to its own development, speeding up the pace of breakthroughs. If we ever hit that tipping point, then things will get weird, and not in a good way.
I think we are as far from it as we were 10 years ago. Or 100 years ago. I think LLM is a deadend technology. Useful, but that won't get anywhere beyond what it is.
But that's the thing, "personally", "I think", etc. Not much of a debate to be had there.
AI making humans obsolete is not really something that causes me any anxiety.
- some jobs will stay with humans even when AI would be better at it. We already see a lot of this with even with pre-AI automatisation. Neither markets nor companies are perfectly efficient
- at the point where AI is better than the average human, half of all humans are still better than AI. For companies or departments built around employing lots of average people the cutover point will be a lot earlier than for shops that aim to employ the best of the best. Social change is inevitable long before the best are out of work
- the actual benchmark for " replacement" is not human vs machine, but human plus machine vs machine alone. But the difference doesn't matter much because efficiency increases still displace workers
- I don't think robots will advance enough to meet this timeline. This is not just a software issue. Humans have an amazing suite of sensors and actuators. Just replicating a human hand is insanely complex. Walking, jumping robots are crude automatons in comparison. We can cover a lot with specialized robots, but we won't replace humans in physical jobs in 20 years
But all of that is assuming a world where research is being done by humans, or by some mix of humans and something like current LLMs. The bottlenecks would ultimately come down to human judgement and human oversight, and that's a significant limiting factor. Plus, you have to push matter around, which takes time, and you have to extract a lot of information out of limited experiences, which LLMs are bad at.
But if someone is reckless and clever enough to build AIs that can completely replace engineers, or that only need humans as hands, then I don't think we can count on robotics remaining intractable for more than a decade or so. In a wide variety of circumstances, it's possible to make do with worse actuators than the human hand, or with specialized actuators. We can already build incredibly precise motors and specialized sensors. The trouble comes with trying to pack enough of them together to replicate the full generality of the human hand. (I have actually helped build task-specific actuators that did quite well with a single motor and a single visual sensor, before.)
So to put my position more precisely: we cannot automate manual labor robotics without having previously automated creative intellectual labor. But conditional on automating creative research, then I expect worryingly rapid advances in robotics.
To be clear, I think that developing fully-general replacements for human intellectual and physical labor would potentially be the biggest disaster in all of human history.