Keep the Robots Out of the Gym(danielmiessler.com)
danielmiessler.com
Keep the Robots Out of the Gym
https://danielmiessler.com/blog/keep-the-robots-out-of-the-gym
43 comments
OP here, yeah, I think that's a really good point.
I feel like the way I'm building this in is a violent maintenance of two extremes.
On one hand, fully merged with AI and acting like we are one being, having it do tons of work for me.
And then on the other hand is like this analog gym where I'm stripped of all my augmentations and tools and connectivity, and I am being quizzed on how good I could do just by myself.
And based on how well I can do in the NAUG scenario, that's what determines what tweaks need to be made to regular AUG workflows to improve my NAUG performance.
Especially for those core identity things that I really care about. Like critical thinking, creating and countering arguments, identifying my own bias, etc.
I think as the tech gets better and better, we'll eventually have an assistant whose job is to make sure that our un-augmented performance is improving, vs. deteriorating. But until then, we have to find a way to work this into the system ourselves.
I feel like the way I'm building this in is a violent maintenance of two extremes.
On one hand, fully merged with AI and acting like we are one being, having it do tons of work for me.
And then on the other hand is like this analog gym where I'm stripped of all my augmentations and tools and connectivity, and I am being quizzed on how good I could do just by myself.
And based on how well I can do in the NAUG scenario, that's what determines what tweaks need to be made to regular AUG workflows to improve my NAUG performance.
Especially for those core identity things that I really care about. Like critical thinking, creating and countering arguments, identifying my own bias, etc.
I think as the tech gets better and better, we'll eventually have an assistant whose job is to make sure that our un-augmented performance is improving, vs. deteriorating. But until then, we have to find a way to work this into the system ourselves.
there could also be an almost chaos-monkey-like approach of cutting off the assistance at indeterminate intervals, so you've got to maintain a baseline of skill / muscle memory to be able to deal with this.
I'm not sure if people would subject themselves to this, but perhaps the market will just serve it to us as it currently does with internet and services sometimes going down :-)
I know for me when this happens, and also when I sometimes do a bit of offline coding in various situations, it feels good to exercise that skill of just writing code from scratch (erm, well, with intellisense) and kind of re-assert that I can do it now we're in tab-autocomplete land most of the time.
But I guess opting into such a scheme would be one-to-one with the type of self determined discipline required to learn anything in the first place anyway, so I could see it happening for those with at least equal motivation to learn X as exist today.
I'm not sure if people would subject themselves to this, but perhaps the market will just serve it to us as it currently does with internet and services sometimes going down :-)
I know for me when this happens, and also when I sometimes do a bit of offline coding in various situations, it feels good to exercise that skill of just writing code from scratch (erm, well, with intellisense) and kind of re-assert that I can do it now we're in tab-autocomplete land most of the time.
But I guess opting into such a scheme would be one-to-one with the type of self determined discipline required to learn anything in the first place anyway, so I could see it happening for those with at least equal motivation to learn X as exist today.
How I see it, LLMs aren't really much different than existing information sources. I can watch video tutorials and lectures all day, but if I don't sit down and practice applying what I see, very little of it will stick long term.
The biggest difference I see is, pre-LLM search, I spent a lot more time looking for a good source for what I was looking for, and I probably picked up some information along the way.
The biggest difference I see is, pre-LLM search, I spent a lot more time looking for a good source for what I was looking for, and I probably picked up some information along the way.
Definitely. We have to find ways to replicate this.
One thing I've noticed is that I've actually learned a lot more code about things I didn't understand before. Just because I built guardrails to make sure that they are built exactly the perfect way that I like them to be built. And then I've watched my AI build it that way dozens of times now. Start to finish. So now I've just seen all the steps so many times that now I understand a lot more than I did before.
This sort of thing is definitely possible, but we have to do it on purpose.
One thing I've noticed is that I've actually learned a lot more code about things I didn't understand before. Just because I built guardrails to make sure that they are built exactly the perfect way that I like them to be built. And then I've watched my AI build it that way dozens of times now. Start to finish. So now I've just seen all the steps so many times that now I understand a lot more than I did before.
This sort of thing is definitely possible, but we have to do it on purpose.
> We'll have to discover where to draw that line in education and training.
I'm not sure we (meaning society as a whole) are going to have enough say to really draw those lines. Individuals will have more of a choice going forward, just like they did when education was democratized via many other technologies. The most that society will probably have a say in is what folks are allowed to pay for as far as credentials go.
What I worry about most is that AI seems like it's going to make the already large have/not divide grow even more.
I'm not sure we (meaning society as a whole) are going to have enough say to really draw those lines. Individuals will have more of a choice going forward, just like they did when education was democratized via many other technologies. The most that society will probably have a say in is what folks are allowed to pay for as far as credentials go.
What I worry about most is that AI seems like it's going to make the already large have/not divide grow even more.
that's actually what I mean by we. As in, different individuals will try different strategies with it, and we the collective will discover what works based on results.
> It is an incredible accelerant in top-down 'theory driven' learning
Is it? People claim this but I really haven't seen any proof that it is true.
Is it? People claim this but I really haven't seen any proof that it is true.
The point that we should do the things we want to be, as in, to have as parts of our identity, is really good insight. Even if the AI can do X well, maybe I want to also be able to do X, therefore I should practice it.
I don't know what those things will be for me, yet, but it's good to have a more specific and directed way to think about which skills I want to keep.
I don't know what those things will be for me, yet, but it's good to have a more specific and directed way to think about which skills I want to keep.
You're really rizzing up the whole "AI can do almost everything better than humans" point. Is there a chance that your investments are causing you to sensationalize things a bit? Because I can promise you, AI can only do better the things that I have absolutely no skill in.
If you need to believe that everyone's only in it for their investment portfolio, for you to sleep well at night, I mean, you do you, but recognize that that's a giant balloon of copium you're huffing.
In fact it's an excellent shorthand. I make it a point to look up what all the AI boosters endlessly posting puff pieces on LinkedIn actually do for a living, and almost without exception it's someone who is directly attempting to profit off of exponential slop machine growth. Or as I like to say:
"The future of money is gold obtained from prospecting." –VP of Shovels, ShovelsCorp.
"The future of money is gold obtained from prospecting." –VP of Shovels, ShovelsCorp.
- using ai effectively is itself a skill that needs training, especially if you're already good at the critical thinking stuff.
- using AI actually most of everything i do anyways, is critical thinking. I'm constantly reviewing the AI's code and finding little places where the AI tried to get away with a shortcut, or started to overarchitect a solution, or both.
- using AI actually most of everything i do anyways, is critical thinking. I'm constantly reviewing the AI's code and finding little places where the AI tried to get away with a shortcut, or started to overarchitect a solution, or both.
If gym is a mindset, hard to separate from the other.
I think that comes down to documenting the mindset as a goal and then using all the AI, scaffolding, and tools available to that system to help you nurture that mindset.
This is a great framework for self-development, but I wonder if the Job vs. Gym analogy is a bit premature. There seems to be a level of Silicon Valley optimism here that assumes AI already surpasses human capability in these creative areas. From my perspective, AI only outperforms in areas where the human hasn't developed a real craft. Is it possible that the current hype is causing us to undervalue the unique quality of human-only output?
How many senior developers do you need to tell you that they've been vastly surpassed (in coding) by AI before you believe them? The creator of Claude Code just said he's not opened an IDE in a while. He was a principal engineer.
And why would you think this would be the only place that'll happen?
I agree there are things they can still do better than AI, but coding isn't one of them.
And why would you think this would be the only place that'll happen?
I agree there are things they can still do better than AI, but coding isn't one of them.
I believe that they believe it, and little else.
Lurking between the lines in arguments about AI writing/code/art is that whether or not an activity is "gym" or "job" is often in the eye of the beholder.
People who never "went to the gym" in a field are all too eager to brush off the entire design space as pure Job that can and should be fully delegated to AI posthaste.
People who never "went to the gym" in a field are all too eager to brush off the entire design space as pure Job that can and should be fully delegated to AI posthaste.
That's why the whole thing has to start with you identifying what you consider core to yourself.
We want machines to do the laundry and clean our house so we have more time to create art and write code. Seems like in our current trajectory, the machines will produce the art and code so we have more time to clean our house and do laundry....
Ah man...that's good.
But maybe both of those are in the category of undesirable things.
And the things we end up with are like art and baking and walking and talking and drinking coffee and such.
Professional Chess is a nice pattern here. A calculator can beat Magnus Carlsen at this point, but Chess is more popular than ever. So it should be ok if AI/Robots are better than us at all the stuff we still decide to do.
But maybe both of those are in the category of undesirable things.
And the things we end up with are like art and baking and walking and talking and drinking coffee and such.
Professional Chess is a nice pattern here. A calculator can beat Magnus Carlsen at this point, but Chess is more popular than ever. So it should be ok if AI/Robots are better than us at all the stuff we still decide to do.
Except Professional Chess, taken to mean players earning a living solely from paid tournament play, is in the low hundreds? Thousands? Meanwhile there are over 20 million 'professional' software developers. There are many things about that single number demographic that I would argue against, but despite that I'm not sure there's ever been a market for any kind of 'professional chess player', yet there is for 'professional software developer' (for some definitions of 'professional' and 'software').
[1] https://evansdata.com/reports/viewRelease.php?reportID=9
[1] https://evansdata.com/reports/viewRelease.php?reportID=9
Making a living from solely paid tournament play is about 20-30 people in the world
Yeah I don't see clankers ever taking over art, music, sports etc. People care in large part about those things because of the human aspect.
I'd love for them to take my job as a programmer though, as that would certainty free up time for me to travel and drink coffee and Guinness.
I'd love for them to take my job as a programmer though, as that would certainty free up time for me to travel and drink coffee and Guinness.
Where would you find the money to do those things with?
If all productive human labor is replaced by AI we have larger problems than where we'll find the money.
I guess we all just need to come up with a shared definition for ‘productive’ first then? Shouldn’t be too difficult.
What if we just use "that which is exchanged for money"?
Hopefully from money I've saved up or a business I started in the meantime.
Maybe cleaning house and doing laundry needs to start being marketed as an art form /s
Ryan Anderson: Are We Having Fun Yet? https://www.instagram.com/itsryandanderson/reel/DRR241Mjr_F/
Very cute. Hardware has very different challenges than software. If you want to get into robotics, it's a growing field with a lot of jobs and money!
Anecdotally LLM vibe coding has sucked ass on the big robotics monorepo I work on
This essay falls under a growing category I call "start off sounding like you agree with critics of GenAI, then subtly attempt to undercut all of the power of their arguments."
The best way to engage with these sorts of articles is to completely ignore all stated advice and move on. There is no "separation of moving things on the job vs. moving things at the gym" when it comes to creative craft, the entire analogy is completely absurd.
The best way to engage with these sorts of articles is to completely ignore all stated advice and move on. There is no "separation of moving things on the job vs. moving things at the gym" when it comes to creative craft, the entire analogy is completely absurd.
> There is no "separation of moving things on the job vs. moving things at the gym" when it comes to creative craft...
- Coming up with names for cities in a role-playing game you're making - Summarizing an idea that you're writing about - Doing research for an article - Brainstorming character names - Creating an aesthetic for a new website for a customer - Etc, etc.
I could go on for days with these examples. And so could any AI.
Pre-2022 ALL these were done 100% by a human.
Now they're not. Now creative people are using AI to help them massively with tons of these. So, yes, the separation needs to happen there as well.
For example, maybe you say, I'll never use AI to help me name characters. Or to come up with plot lines. Or whatever.
That's a Gym vs. Job distinction.
- Coming up with names for cities in a role-playing game you're making - Summarizing an idea that you're writing about - Doing research for an article - Brainstorming character names - Creating an aesthetic for a new website for a customer - Etc, etc.
I could go on for days with these examples. And so could any AI.
Pre-2022 ALL these were done 100% by a human.
Now they're not. Now creative people are using AI to help them massively with tons of these. So, yes, the separation needs to happen there as well.
For example, maybe you say, I'll never use AI to help me name characters. Or to come up with plot lines. Or whatever.
That's a Gym vs. Job distinction.
"Now they're not." ???
For me, there isn't the slightest difference from before 2022 and after 2022, since I continually choose to boycott genAI services, and as an activist in the Pro-Craft movement, encourage others to do the same.
For me, there isn't the slightest difference from before 2022 and after 2022, since I continually choose to boycott genAI services, and as an activist in the Pro-Craft movement, encourage others to do the same.
Yes, I could tell. : )
You do realize that the entire point of the post is to be cautious of AI, right?
And to make sure that you have your own personal goals separate from it, and that if you're getting help from it you need to make sure it's in line with those goals.
Right?
What about that do you disagree with?
And to make sure that you have your own personal goals separate from it, and that if you're getting help from it you need to make sure it's in line with those goals.
Right?
What about that do you disagree with?
I disagree with your insinuation that being creative with your mind as a human being is a form of recreation or personal growth or whatever (working out at the gym) whereas "creativity" on behalf of work projects gets done by Big Tech's generative algorithms. It becomes even more confusing because then later on you say "I inevitably will use AI to do many of these Gym tasks" …huh? So now even the creative "gym tasks" are done by AI too? Apparently our personal goals will shrink to a freakishly tiny little list by the end zone of your thought experiment!
> [AI/LLM writes] why I made the decisions I did
When one thinks about human decision making, there are at least two classes of decisions:
1. decisions made with our "fast" minds: ducking out of the way of an incoming object, turning around when someone calls our name ... a whole host of decisions made without much if any conscious attention, and that if you asked the human who made those decisions you wouldn't get much useful information about.
2. decisions made with our "slow" minds: deciding which of 3 gifts to get for Aunt Mary, choosing to give a hug to our cousin, deciding to double the chile in the recipe we're cooking ... a whole host of decisions that require conscious reasoning, and if you asked the human who made those decisions you would get a reasonably coherent, explanatory logic chain.
When considering why an LLM "made the decisions that it did", it seems important to understand whether those decisions are closer to type 1 or type 2. If the LLM arrived at them the way we arrive at a type 1 decision, it is not clear that an explanation of why is of much value. If an LLM arrived at them the way we arrive at a type 2 decision, the explanation might be fairly interesting and valuable.
When one thinks about human decision making, there are at least two classes of decisions:
1. decisions made with our "fast" minds: ducking out of the way of an incoming object, turning around when someone calls our name ... a whole host of decisions made without much if any conscious attention, and that if you asked the human who made those decisions you wouldn't get much useful information about.
2. decisions made with our "slow" minds: deciding which of 3 gifts to get for Aunt Mary, choosing to give a hug to our cousin, deciding to double the chile in the recipe we're cooking ... a whole host of decisions that require conscious reasoning, and if you asked the human who made those decisions you would get a reasonably coherent, explanatory logic chain.
When considering why an LLM "made the decisions that it did", it seems important to understand whether those decisions are closer to type 1 or type 2. If the LLM arrived at them the way we arrive at a type 1 decision, it is not clear that an explanation of why is of much value. If an LLM arrived at them the way we arrive at a type 2 decision, the explanation might be fairly interesting and valuable.
Does it really matter how the LLM got to a (correct) conclusion?
As long as the explanation is sound as well and I can follow it, I don't really care if the internal process looked quite different, as long as it's not outright deceptive.
As long as the explanation is sound as well and I can follow it, I don't really care if the internal process looked quite different, as long as it's not outright deceptive.
I'm just quoting the author of TFA, who did in fact appear to want periodic explanations of how their "agent" arrived at its decisions.
It is an incredible accelerant in top-down 'theory driven' learning, which is objectively good, I think we can all agree. Like, it's a better world having that than not having it. But at the same time there's a tension between that and the sort of bottom-up practice-driven learning that's pretty inarguably required for mastery.
Perhaps the answer is as mundane as one must simply do both, and failing to do both will just result in... failure to learn properly. Kind of as it is today except today there's often no truly accessible / convenient top-down option at all therefore it's not a question anyone thinks about.