Linkies
Learn in public
A great post with practical advice.
The fastest way to learn is in public, which means creating and sharing content related to what you are learning.
It benefits yourself and others, as you document your progress, get feedback, find mentors, and build an audience + overcome impostor syndrome, gain confidence, and showcase your skills.
You can learn in public by writing blogs, making videos, answering questions, teaching workshops, cloning projects, etc.
Don't be afraid of being wrong or criticized, but use it as an opportunity to improve and prove yourself.
Motivation is overvalued
Classic Atomic Habits piece.
Motivation and talent are often overvalued. Environment matters more in shaping human behavior and outcomes.
To design a better environment for success, you can automate good decisions, get in the flow of your normal behaviors, and subtract the negative influences ⇒ nudge yourself.
Solomon's paradox in life
Solomon was incredibly wise concerning others, and an oft idiot with personal affairs.
Solomon’s paradox is the phenomenon that we are better at advising others than to ourselves.
There are two types of wisdom: general wisdom (interpersonal) and personal wisdom (intrapersonal).
To improve our personal wisdom, we can use self-distancing techniques such as self-talk (Ask two questions, to start: “Why are you doing that?” and “What can you do to help?”), journaling (not CBT-style, just plain journaling + possibly transcribing voice notes), or identifying with someone else (ask a friend or relative who they think you resemble and learn from their life experiences).
Berserker hypothesis
is a proposed resolution to the Fermi Paradox, which asks why we have not detected any signs of extraterrestrial intelligence (ETI) in the universe.
- The hypothesis suggests that ETIs have been wiped out by self-replicating machines (von Neumann probes) that were either programmed or went rogue to destroy all life they encounter.
AI jobs
So, according to someone, AI is going to steal our jobs. Setting aside con, blow- and rim- (don't google that, I've just made a slide deck on kinks, I'm a professional) ones - let's speculate
Physical
Let's be honest - while you can use an AI for some intern/junior-level jobs (web research, programming) right now, the time for it doing even home chores is yet to come: a study had estimated ~40% of home chores being automated in a decade. Yeah-yeah, we've entered the phase of self-improvement, yada yada - but still.
Let's remember the funny vids of PaLM-E for instance. Boston Dynamics and Halodi are the epitome of what'd I see robots fit: freeing people from stressful, dangerous and mundane/routine jobs. What are we gonna do with all the displaced? I'd guess that's a cognitive bias often mentioned: as a whole, humanity will have more to do, starting with progress/education and, say, art. Or...exploring the world?
Emotional
You can fake empathy and anthropomorphize the robot, that's what ELIZA effect is about. And yeah, a guy recently took his life after an AI convinced him to do so - what did I say about con jobs, again? Forget it, after getting good open-sourced LLMs we're focked, mates 🤡
However, barring that, we've got genuine affection, touch, and attention - the things that won't be reproduced for a long time. I can presume that, after machines become sufficiently advanced and pass the uncanny valley, two camps will emerge:
the ones who embrace and are in for feeling genuine emotions towards a robot;
the 'you don't understand, it won't replace us' old (?) believers.
I may be genuinely wrong, and the robots will learn to express, feel or fake + we'll all get into it (see what I did here). Anyway, we're in for some biig ethical debates and hard decisions for some to make.
This bad boy can treat paedophilia in a day!
Teleogenesis vs initiative
Again - it's not the 'I should do X because I've got a higher-level task X[n+1] and it's the highest-priority task here', but 'let's do X' or 'LFG'. I may be mistaken, but human initiative/motivation is unique in the sense that it's not just about completing a task but also about the drive to do it and the satisfaction of doing it well. This is something that machines cannot replicate (yet).
Can a machine replicate Ryan Gosling? Can it express a negative amount of emotions? Can it get critics' appraisal for its acting performance?
"AI" vs AI
Another thing to consider is the flood of "do ... with AI" products. Let's be honest, most are convenient yet fancy wrappers over, most often, OpenAI's API.
Another labeling galore incoming:
Prompt shops
Basically, 90% of current Product Hunt. Yes, they make our lives easier and leverage the latest scientific advances of AI, blah blah blah. But:
they're usually dependent on vendor lock-in -> API changes;
the results are potentially dispersed with each foundational model update;
Something like that. Maybe, I'm a bad 40-second-coder.
- weak moat/defensibility
No defensibility in either brand, scale, embedding or (sometimes) network effects. No moat - such a business may get astray (or in an ashtray) with, say, MS Office's Copilot announcement.
'Real' AI products
Usually 'dependent' on a team of engineers/scientists. Usually not that loud in their announcements (unless it's some big-ass company running for an AGI, of course...).
Usual AI engineer
Welcome to Teleogenic/YAWN/Boi Diaries❣️
You can find the other blogs I try to cross-post to: