Signal

the new skills that are getting to be more important are giving feedback. you give feedback to ai’s output to get what you want. taking a take. having a position/opinion on some aspect. sequencing. knowing what to ask first. taste. differentiating what’s good from a mediocre output. explaining things. the better you explain/give context the better you get the output. (i might put this in the bracket of context engineering) knowing when to stop. over iterating kills the edges. and none of them were called skills before. but now ai made them move to the front.

writing could be put in two aspects. stating intention & expressing something. ai writing makes sense when stating the intent itself is enough like emails, work reports, leave letters, refund requests, notes for rescheduling appointments and so writing it ourselves is lot better when we’re adding a touch of emotion and want the reader to feel what we feel. like personal messages, apology messages, journals, thought descriptions, resignations, first dms, farewell messages and so the lines might blur sometimes but holding onto them would yield a personality to our name.

on a visceral sense, i think these could be the new types of jobs. context engineering agents orchestrator human-ai interface designer (ux not ui) integration specialist ai compliance officer and prompt engineering is just language. you need to know the language to talk to ai. it’s not a job type.

science is what the world is of and commerce is what humans operate on. once someone understands the layer of how things work, then the operational layer kinda becomes lil easy. and so we see the pattern of engineering and then mba. you kinda develop a thinking mental model first before the operations. doing it reverse is usually hard but maybe cuz of ai, it’s kinda not as hard as it was but still, if i had to push a kid to what to learn, i’d push towards a science first education.

i believe there are two things to do in life. serve a purpose & raise a good kid. i mean what’s life anyway get in, do it and get out. period. & enjoy the time doing it. there are so many nuances ofc but end of the day it all just is that.

ai is the greatest test for human inhabitability. the human adaptation usually takes shit ton of time, sometimes days, weeks, months, years and even generations. we generally take longer to adapt to things as we’re usually comfortable at what we’re already been doing or we’d have a hard belief. like, the process of ‘oh that’s interesting, let me give that a try…oh shit that works.. let me start using it too’ this part takes time. sometimes a lot. cuz in between the process, we flinch to old techniques or get carried away or something doesn’t work and give up and every step is consuming time & it adds up. and with ai, we’re on a hyper adaptation curve, as things change everyday, the things/techniques we clung onto, gotta go & start to adapt new ones almost everyday/week.

reading about evolution is crazy interesting, one of it is, solving a problem biologically takes thousands of years, and humans (sapiens) slowly learnt to redirect that natural process by starting to use tools eg: sharp rocks, arrows, fire, farming.. and using a tool compressed the time lines of solving a problem externally for which biology also slowly adopted and pushed humans into advancements at an incredible rate. now, the progression of ai in the last few years gave us soo many tools that if biology alone had to bring us those, it would take even longer than the whole universe’s existence. (13.8 billion yrs) for comparison, in tool development, going from using sharp rocks as a tool to attaching a wood piece to it took around 10,000 years.

one of the things i’ve started really late is documenting & categorizing. or putting things in one place in accordance. either they are thoughts, music, to-do lists, to read pieces, goals, people, project ideas, conversation insights, things i learned, references for later, inspiration that hit at 2am, patterns i notice, questions i want to explore and so structure is just easy to look back at and act upon than figuring out from scratch every time.

people approach ai in different ways like some treat it as a friend, some treat it as stranger with knowledge, some treat it as glorified google search i like to treat it as friend. like i address it with ‘yo, bro, dude, g, man’ and so. just like the suffix’s i use while texting my friends. i felt it like if you start it with a suffix, you tend to express well & ask the question even better than if you’re baldly just framing the question.

movies need to have some kind of propaganda. without it, it’s just hollow. and a bad movie is either a good propaganda shown poorly or ideologically misaligned with us.

as you orchestrate more ai agents around you, pretty much your work becomes something like, you wake up in the morning and the first thing you’d do is to read a bunch of reports. (your agents doing a bunch of work & telling you what’s up now) then you get to work, talk to lots of people, make a few decisions (maybe actually a lot) and at the end of the day, you’d tweak the way you get the reports next day.

i had this instinct early in the morning while waking up that "evolutions happen cuz of technology/new methods" so after freshening up, i sat down with claude & chatgpt to understand it better. i then got little rough sense of the stages of evolution and got into the part of exploring different types of histories. like history of energy, information, tools & coordination. then i made chatgpt do a deep research about energy, it went through 73 papers and wrote a report. i then gave it to notebooklm and turned it into a podcast and started listening while going out. the whole thing has happened from the time that i got up & finishing drinking my tea. now, i might or might not remember all this in a week or later but atleast i think i'll remember the larger patterns. we're obviously on the brink of a new evolution but for the first time it's happening faster than ever.

everyone at somepoint has to go thru the phase of learning how to work/live/think with ai. it’d be either for you to use it or to teach your kids. it’s unescapable as of our existence in this current time frame. simply put, do or die. and the learning phase will not be an overnighter.

i think there’ll be too many people loosing shit with the autonomous agents having too much of access/control of you. but they’ll hardly comeout and admit. cuz of two reasons. 1. there’s no undo option 2. admitting that you got fucked cuz of you giving access to it reveals soo much of you, or maybe it’s a genuine mistake but rest of the world automatically judges. and to avoid it, you will not admit or come out. and my guess is this market alone will touch a few billion by the end of this decade.

strange that zooming in shows us more, but zooming out helps us see better.

there are three ways to consume content/information text (reading words) video (moving images with audio) audio (sound only) with text, you move slow. you read it, you process it, and then you come to a conclusion/opinion. your brain is doing lot of work. you burn a lot of energy. with audio, you're listening, you're imagining/picturing and constructing the scene in your head, your brain is constantly filling in the gaps. again, you burn a lot of energy. with video, the information comes to you semi processed. the visuals, tone, emotion are already set. you're pretty much only reacting. your brain doesn't have to work as hard. you burn lot less energy. (not the podcasts/talk shows/interviews filmed on camera. they're just audio with a face. i mean actual video, where the moving image is carrying information.) and in generally, we humans usually pick the easy route. shortcuts. just so we consume less energy. it's in our dna. but to get any deeper insight, you gotta burn energy. just like an energy burning body stays fit, the same applies to the brain.

we got to the times of creating software on a whim like the websites, apps, trackers, presentation wares, dashboards, automations, tiny games, and so so when the cost of creation drops low, the value for ‘why’ shoots up. but when a single person is creating a crazy number of types of software, it then comes down to how much of a value the creator is scooping of the creation. as in, it’ll be about the ability to generate the right things at the right time for the right reasons. and i think that’s the real personal software. maybe instead of saying ‘i use x or y software, we’d be saying i use this x person’s software instead’ maybe.

our digital real estate is mostly built on scroll & consume phenomenon. with ai video apps on the rise, doom scrolling traps are getting multiplied and might even get stickier, but for our attention. we gotta be real careful about why we’re scrolling and what we’re looking for in the feed, or else we fall into these traps easy.

getting started requires massive effort, but once you’re on track, continuation becomes nearly automatic.

one interesting pattern from our evolution was, writing was invented in 3400 bce and the argument was “if we’re offloading knowledge on to papyrus(paper) we’ll stop remembering things” then, printing press was invented in 1440 and the argument was “if mass produced books gets in hands of unqualified, then we’ll have people with misinterpreted knowledge” then, calculators/computers became widespread in 1970s and the argument was “if we’re outsourcing math skills, then we’d be worse at math skills” then, internet became publicly accessible in 1991 and the argument was “if we have access to everything instantly then we’d search for every answer and it would make us shallow” and now the ai. which became publicly accessible in 2022 and the argument is “if ai can write, reason, analyze for us, we’ll lose the ability to think deeply or maybe even think” every time, the invention moved humanity forward. not just externally but also how our brains biologically work.