hello again, to the seven or eight of you that read these!

i've been doing a lot of writing (in private) over the last 2 weeks so figured i should get back into weekly blogging on here again.

as i've been working on my distributed training report - which is very good btw - i've found myself using deep research to help me dig deeper into more niche topics.

i subscribed to openAI's pro membership maybe a month ago, and while i didnt make use of the 200 deep research queries in that time, ill probably burn through them by the end of this week.

people like to point out the flaws of AI and how it isn't comparable to the best humans, and i always hate seeing these tweets go semi-viral.

it isn't useful to compare AI to the best legal assistant, best investment banker, or best scientist - compared to google search, AI makes my life 1000x easier, and deep research multiplied that effect in a very short amount of time.

how do i use it?

i've messed around with deep research to generate in-depth reports about niche topics like the shipping or aluminum industry, backgrounds and analysis of somewhat low profile historical figures, major trends within niche sectors of tech, and the list goes on.

through all of these examples i was able to learn many new things without doing any work (outside of two kinda detailed prompts) and it saved me maybe hundreds of hours of work.

i understand $200/month is a lot of money, but how much do you value your time? if you're working in any industry that involves a lot of reading, sifting through various sources, writing, crafting insights, or whatever else fits into that list, deep research is 100% worth it.

deep research's biggest strength isn't in its analytical abilities or its writing prowess (it isn't very good at either) but the speed it can dig through fairly under the radar resources across the internet and pull all of this info together into a report that's more often than not exactly what i wanted.

in the past, when i wanted to learn more about an industry i would have to spend hours and hours just running basic google searches, scanning through reddit reply chains, watching YouTube videos, or trying to find a PDF of some book that's been out of print for three decades.

now that a tool like deep research is available, there really isn't anything stopping me from gaining entry-mid level awareness of any topic i can think of. you can even use it for really simple stuff relevant to your daily life, like asking it for a report on the interactions between supplements/vitamins you take, studies and analysis of what workout routines lead to x y or z desired outcome, and so on.

it isn't even a requirement that you need to do some type of knowledge work to use deep research, because it's pretty cool to just ask it about any interest of yours and read a quality report in less than 7-10 minutes.

there are a few alternatives that i haven't tried, like the perplexity research tool, the new manus agent/assistant, open source alternatives, and even some anecdotes ive heard about claude 3.7's capabilities as a deep researcher.

im sure all of these offer the exact same quality of writing, but im not certain any (except for maybe perplexity) can dig for sources as well as openAI's deep research can.

it isn't like im taking these reports and posting them online, so i really dont care much about the quality of writing. this is much easier and informative than reading a few articles online or scanning a wikipedia page.

even if deep research didn't improve from here, i dont see why id ever unsubscribe unless prices were dramatically raised. as far as utility goes, i see something like deep research in the same way programmers view cursor - it's additive to your workflow, and obviously shouldn't be a crutch for actual experience or code writing (in this case, writing writing).

go try it out, it's a lot of fun, i think it's free too for non-pro accounts.