I know that the common refrain is “think of yourself as a manager now” but I’ve actually taken the opposite approach and have been telling anyone I train the same.
Diving deeper into technical understanding makes more sense to me at this point both as a way to make yourself more useful in the age of AI and also to use AI more effectively.
I regularly tell the kids to grab a text book on a subject that interests them and I do the same.
I’m willing to bet deep understanding is going to become a commodity soon.
I've been a manager for some years. Without sounding too up myself, what made me good at management was the fact that I knew how to do the things I asked people to do. It allowed me to critically evaluate the work, set realistic timelines, and champion the contributions up the chain. I also maintained the ability to perform this work (by design and happenstance sometimes, since we could be understaffed due to sickness etc).
The good managers I saw also had this ability. The managers I didn't enjoy working with / saw struggle were ones who either never had that ability, or had lost it a long time ago.
Note I'm speaking about line management. There's a world of difference between managing a dozen or so individual contributors vs managing senior managers / directors. However, managing LLMs is analogous to line management.
> There's a world of difference between managing a dozen or so individual contributors vs managing senior managers / directors.
I agree, but in my opinion, your point about needing to know how things work still holds. It's less relevant as you move up to managing managers and larger orgs, but it never goes to zero if you're going to be successful.
I think it holds true throughout if you scope it to knowledge about your reports’ domain. If you’re managing managers you should have a good handle on what a manager does and different ways to do it.
There's two huge aspects of management that gets diluted when it's compared to "managing" an LLM:
- labor laws. You need to know those in the country where the employee has their contract, because this will restrict what you can (and must) do in your managerial role
- Empathy. Managing humans is extremely messy and requires a ton of EQ to do well (parents have first-hand experience of this).
So, i.e: comparing being a manager to real humans vs instructing a next-token generator dilutes the skill of being a good manager and leader.
It is easier than ever to learn difficult concepts. It is also easier than ever to produce things that used to require understanding of those concepts without them. Discipline and drive to use these powerful new tools patiently and with purpose is what is required.
Do we actually have any measurements of if AI helps you learn something better?
Because it is one thing to feel like you have learned something and another thing to have actually learned it well enough to put it in use. There are so many YouTube tutorials on all kinds of subject that will make you feel like an expert in the field after watching, and then you start doing the thing they have supposedly taught you and you can't remember a darn thing, because your "learning" was only ever skin deep and never meaningfully tested or reinforced.
> Do we actually have any measurements of if AI helps you learn something better?
To me this is intuitively true based on anecdata. 2 examples
1) learning spanish - when I hear or read a phrase I don't know, I type it into the LLM and I learn a new word/phrase. Sure, I could have cracked open a spanish language dictionary, but tbh, I wasn't going to do that. Not to mention that dictionaries are translating word by word and not phrase by phrase.
2) growing vegetables in the garden. I literally watched YT tutorials and did what they did, and now I have vegetables that I didn't before. Yes, I could have probably could have gotten this from a botany book, but once again, I probably was not going to do this. I also was trouble shooting a lot in gemini
Here are the problems I traversed:
- How much water should I give these per day? what's the watering schedule?
- how much sun vs shade?
- when do I move seedling to the ground outside?
- Is trimming good? which parts and when do I trim?
- [take a picture of weird growth] - Is this a disease on the plant? Or part of the plant naturally?
I also have one anecdote - I (think) I understood basics of quantum mechanics intuitively for the first time. Of course it's probably superficial and maybe not 100% correct, but it is better than every time I tried to understand it before.
It definitely wasn't a single prompt, but two hours of back and forth, with a lot of time spent thinking (me, not LLM) in between. There were multiple times where I misunderstood something, so if I just read a book I'd probably get stuck many times.
Do you retain that information? Or are you just constantly relying on having access to an llm to re-look things up. The amount of information I retain when using llms to learn is far less than other methods, not sure why, you'd think it would be just as effective as reading a book, but it's definitely not. Probably for the same reason that learning math and always using a calculator does not make you great at math.
I've wasted a ton of my life already trying to make llms work for learning over the last few years, I'm especially bitter about it. I think this technology is a scam made to make us reliant on a think-for-me machines.
I definitely retain it. In the case of the gardening - I pretty much immediately go outside and practice it.
In other words, how I get the info doesn’t matter (book, video, llm) because I’m locking in the skill by doing the thing in the garden.
And sometimes it is kinda wrong/suboptimal, but not more than e.g. asking a friend for gardening tips. Usually this is based on lack of context, like eg im growing cherry tomatoes and the AI assumed I had full size tomatoes
I dont think you are getting the topic - sure if knowledge gleamed from llms is immediately practiced repeatedly, it stays, why would anybody question that.
Now add 30 or 50 or 100 other topics you dip into on any given day. No way in hell you do same technique of repetition till you learn it. This is gist of problem to me. Most students these days fall into this category and the amount of time/energy to properly learn something has stayed the same or even increased (due to distractions and pushing through mental barrier of 'I know it well enough')
> Now add 30 or 50 or 100 other topics you dip into on any given day.
In a world without YouTube/LLMs, you weren't going to learn these things anyway. Today, you also don't learn them. So how can you argue that something is lost?
> the amount of time/energy to properly learn something has stayed the same or even increased
The whole point of my post is that LLMs, YouTube, and other internet resources which are easy to access have lowered activation energy to where learning something is at least possible.
I can tell you for sure I would never even try to learn a second language if I had to go to a class physically, or carry a book with me in my pocket to translate as I traverse the real world.
I think it depends on how much time you spend learning something and "engaging" with LLM.
I have many times asked it something I was slightly curious about, got the answer after the first or 2nd-3rd prompt, spent 3 minutes in total and forgot it after 15 minutes probably.
But a few times I've spent an hour or more on a topic, asking many questions, thinking between responses, and I actually learned something.
For me it depends, if I simply take the answer from the LLM it goes straight into the shortest memory neurons my brain has.
Now, if I go back and forth with the LLM to say, taking the language learning example, to explore the etymology of the word (which for me is far more interesting than the translation itself), then I learn a ton more.
I'd add that this is also how it used to work with Google search too: there's been facts that would show up as the first result, then you'd just forget it 10 minutes later and have to re-google it. I know this happened to me many times.
For better or worse, there's less friction now for seeing an answer to a question you have. You can ask in more arbitrary ways than Google required, and something will still come up. For looking up factoids, it's much faster. For picking up more complex topics, I'd say it's more or less the same, because you still have to spend time ruminating on the topic.
There is some evidence to support the idea. However, "learning" studies can be a little disappointing since they usually take place over a short period of time. This one had a week-long timescale:
I think there's a bigger question that would be hard to answer: what does a consistent long term diet of videos and quick answers do to a person's mind/understanding of the world?
Also, a lot of the published research is a year delayed - most I've read recently are using ChatGPT 4 or 4o. When it comes to giving high quality feedback, I think there was an inflection point around Opus 4.6.
I wonder if it'd be more effective to teach how to critically use AI than to try to forward people to textbooks.
For myself, I have found that I am better able to learn new topics than ever before because being able to have a conversation with a moderately competent but sometimes catastrophically wrong AI about any new subject is actually the perfect mix of helpful and unhelpful for learning.
I use a loop along these lines:
* Ask a question
* Get an answer
* Be skeptical of the answer
* Investigate/reason about the answer
* Critique the answer
* Rinse and repeat
This kind of loop is far more useful to me than any textbook ever has been, because a textbook just drips information into my head. It's more likely to be accurate, but not guaranteed, and it doesn't encourage me to actually engage with the material in the way that a wrong AI answer does.
That’s the beauty, you can do both. In practice I usually just let the AI know what chapter I’m on and then ask questions or have it ask me questions based on the chapter.
> the common refrain is “think of yourself as a manager now”
One of the many reasons I'm determined to remain a luddite wrt AI. I hate the idea of being a manager and have refused promotions to avoid it in the past. I don't want to manage automatons any more than I want to manage people. I want to do stuff, not manage.
I find a lot of people are actually happy that they can offload mundane parts of job to automatons and to be really busy with the interesting parts.
It is not like you have to give them fun parts of the job.
While being a team lead for people I actually have to do the opposite. Get interesting parts to hand over to people to keep them happy and pick mundane not interesting boilerplate myself.
I keep hearing my this very much, so I have a question: what fun parts are there left for me?
It feels like this message of “offloading mundane tasks so you will have more time to do fun parts” is being pushed so hard, but in reality the opposite is happening. Fun parts are being offloaded while I’m left doing the soul sucking and mundane parts
I need a bash script for example. I don’t like dealing with bash or debugging scripts written by others.
I have that part covered by AI so I can get multiple servers up and running and then eventually fix some config or networking here and there which I like.
I love churning out crazy Bash scripts, though. The first scripts I wrote sucked, as it always goes with a new language, but Bash is an awesome swiss army knife now that I know how to really push it. It sucks that doing this by hand now is so heavily discouraged because so many people see this as drudge work (that they never cared to learn).
I dislike the tooling and debugging experience that is not existing compared to let’s say Python. With Python I have whole IDE which help me set breakpoints and inspect variable state. With bash I have to pipe output to text file and inspect that. Feels like working with Stone Age tooling.
I'll take this as an uninvited opportunity to plug nushell. Ive never got on with bash, it feels like too much of the friction is incidental to the problem, whereas with nushell I find I get to the fun intrinsic difficulty of a problem. YMMV but give it a try if you haven't before
I'm a bit odd, it would seem, in what I find enjoyable and not.
I've happy to take the tasks like SQL optimization which other would run away from, for instance, or the bash trickery mentioned elsewhere. People don't believe me, I've had looks of surprise and derision when “if you take on this role, you can give X to others and not have to worry about it” is responded to with “I like doing X, can't you ask one of the others to take on the other role and let me stay as I am?”…
It is increasingly looking like I have the choice: change to a role (automaton farmer) that I expect to hate, or leave dev, or maybe leave tech entirely and go wait tables or something. Those tables are starting to look very attractive.
I too give juniors the advice to crack open $textbook. It’s just painful to see the complex things they’ve created with horrible performance and no cohesive data model because they don’t have the requisite academic foundation to hand code the same thing given unlimited time
Managers, good technical ones anyway, also dive deeper into technical understanding. They don't do it to write more code, they do it to figure out what the developer is saying to ensure the developers aren't blowing smoke.
Your argument in that quote is completely unclear. Why do you believe that deep understanding is going to become a commodity soon? How would that happen?
Are you really diving into technicals just by asking "sooo what did you do?" to your AI? Without document crawling, debugging and pulling your hair out, how much of it is really a deep dive? I feel like all that effort that goes into generating a mental image from bits and pieces is gone. I'm just grudgingly happy I went through all that before humanity retired.
The system does not reward deep understanding. It's too slow.
An analogy that I think helps describe how I feel about it all:
AI is like having cheat codes in a video game. You don't have to try hard and develop a deep understanding of how the game works. You aren't challenged anymore and it doesn't feel like you earned winning the game.
It doesn't matter because nobody cares. Businesses truly do not care. You are a cog, a means to an end. It's only about "winning". Now it's no longer fun to play the game.
> It doesn't matter because nobody cares. Businesses truly do not care. You are a cog, a means to an end. It's only about "winning".
A cynical person would tell you that nothing has changed; this was always the way it was. AI didn't change anything in this regard. Maybe it just made it more apparent to you?
I couldn’t agree more. It seems AI is very good at middle of the distribution tasks where it has access to a lot of training data - enough to be highly reliable at that task.
Which means as a human your only added value is on the edge of the distribution. Which means you need to be learning and doing more complex, deep topics.
Diving deeper into technical understanding makes more sense to me at this point both as a way to make yourself more useful in the age of AI and also to use AI more effectively.
I regularly tell the kids to grab a text book on a subject that interests them and I do the same.
I’m willing to bet deep understanding is going to become a commodity soon.