How Others Use AI WITHOUT Losing Their Mind

0:00 Can you use AI without losing your soul or your mind? It's undeniable that internet is ruined with AI slop. But the question is, are all the users of AI bad? Originally, I wrote an essay

0:13 regarding this topic. And then I thought, instead of me talking about how I use AI, how about we do a rapid fire of real people leveraging AI to save their time using AI as data processing,

0:26 what it's good for, not AI images, not content, for some vague likes, and in some cases, you will see actually save their livelihood. When February hit and everything fell apart, I still had

0:39 income. I could still pay for medical care I needed. The job AI helped me get and excel at became the financial foundation that kept us stable during the crisis. This is the use cases and

0:51 stories that I see every day. Those are not on social media are those people who are just building for their livelihood. They are not going to make YouTube videos about it. What gets amplified on

1:03 media is loud and polarized. Sheila's story is absolutely incredible. Sheila herself and her mom went through insane health journey and is still going through. So in this story AI and

1:17 learning to build automations and helping with her own job actually saved their life. Then she could not work. The crisis hit. I passed my automation studies. I didn't have bandwidth to

1:28 learn anything new. But the automations I already built, they became my lifeline. Work week dropped from 40 hours to 15 hours. Same output, same quality. My client is still happy, but I

1:40 gain 25 hours a week. So, what did Chila do? Back in 2023, Chajip was helpful, but very difficult to tame. I studied prompt engineering, but it was Goddess Tricks that helped me achieve the

1:51 writing tools I needed. The newsletter I built got nominated as top 10 newsletters in fintech by Beehive. My client has been happy and consistent performance from 2024 until now. This

2:05 matters because when February hit and everything fell apart, I still had income. I could still pay for medical care I needed. The job AI helped me get and excel at became the financial

2:16 foundation that kept us stable during the crisis. Sheila just became claude certified architect as well just few days ago and it's just privileged to witness her journey and responsible use

2:29 of AI and the content that she creates. Well, being in top 10 newsletter for the client, I think speaks for itself. Okay, Dylan, clot code now does the task that I used to hate. I was manually reading

2:42 PDFs and books, documents, writing detailed notes, spending maybe one to two hours on 50 to 100page PDFs. Now, I can just read the PDF. So, this is important. Dale is still reading PDF.

2:54 and I drop it into the folder, run the clot skill I created, and boom, detailed summary, notes, frameworks, and models are all there using the original document pages as citations in case I

3:07 need to deep dive Vicki style. It might take me 15 to 30 minutes to read the PDF. Now, conservatively, 30 to 40 minutes of study time saved per PDF. This is so relevant to me as well

3:21 because I read um I read research papers myself and I use remarkable to highlight and uh make my own notes and then I just send it to my email and my AI assistant picks up and can extract summaries notes

3:35 whatever references or just see what caught my attention and I can chat about it further. Okay, vines cloud code can now do the prospect research I was previously doing manually. I'm no longer

3:47 spending most of my time collecting information. I can spend more time reviewing it, deciding what matters and thinking about how I would actually approach the business. So again, AI is

3:58 perfect for data gathering, accumulating the data, processing the data, but you are sitting in the driver's seat and spending more time to actually thinking about that information on the potential

4:08 business lead that been captured. Mike week two commitment for Mike was migrating clients side of WordPress onto static HTML on Cloudflare. So before default Claude wants to improve

4:22 everything. Yeah, AI will rewrite the whole thing with its AI slop writing. It renames URLs that carry years of ranking, drags across boilerplate text that has never true for the content and

4:35 so on. Now after one command, it captures DNS before touching anything, checks every existing URL, moves the client's own copy untouched, writes explicit 300 ones, and ask which page

4:48 are sensitive before going near them. The rules that used to live in my head are now checks that block commits when a site breaks one. And this is real work. Sasha, my assistant now does a daily

5:01 check and sweep of my download folder on my computer and moves them into a subfolders of another folder I have created. It does this at scheduled time each day and then reports to me via

5:13 Telegram what has been done and any new subfolders that had to be created. It also keeps me updated on items on my Google calendar. Next step is to connecting to my war calendar. I don't

5:25 know about you but this use case my downloads folder used to be just absolute mess actually my whole computer all the files would be a mess and that AI can manage your files and not only

5:36 manage and move in the subfolders but actually indexing them so that it knows where is what and how things are organized as you notice some of these posts start with the same sentence cloud

5:47 code does the task I used to hate so in our 8week accelerator the first week we prompt members to think about something like a one workflow very simple input output that they hate. Not something

6:01 that they want to create, not the whole giant automation system. Just one little thing that you hate and if it was gone, you would not miss doing it by yourself. That's the whole point, right? Okay.

6:13 Every video worth keeping me being at my desk. Open my second brain, paste the URL into ingest panel, wait for the transcription, run the transform, save the receipt. about 5 minutes of

6:23 babysitting each time. The real cost was not 5 minutes though. Links I found on my phone just piled up and never went back to them. So I feel this use case so much because I'm on social media, right?

6:38 So I save videos and they used to go into some place or take screenshots, right? And then it goes to this black hole it never comes back to. So how does it look now from the phone? I mind you

6:51 from the phone. I share the link from my phone to a dedicated inbox and I forget about it. A watcher picks it up, pulls the audio, runs whisper picks and whisperers transcription service, picks

7:03 the right transform prompt and saves a finished receipt into my second brain tagged email link. This is fantastic. I shared with Mike that what I do now anything that I save from social media I

7:15 save it as a links to Apple notes folder and then from the Apple notes folder my AI assistant picks up those links it's like automated trigger and basically puts it in a database and then I can

7:27 surface and ask questions and it can surface those exact things that I saved. Mark is in finance so this is like obvious tool that he built. So again, week one commitment. Even though I would

7:37 say this commitment has been a little bit big for week one. His AI automatically reviews investment portfolio by analyzing each stock I own, summarizing relevant news and market

7:49 sentiment and identifying buy, hold, sell or trim opportunities and evaluating my overall portfolio for concentration overlap or rebalancing opportunities. Now pay attention to this

8:00 one. Mark is not saying that he build a financial assistant that buys stocks or trades or makes a decisions. This assistant helps him. He doesn't have to do manual research on his every single

8:14 position and just sit drowning in X. That is the researcher assistant consolidating information and then Mark is going through it by himself. And mind you also Mark is in finance. He is a

8:27 domain expert. He would be able very quickly to notice if there is some discrepancies and how to get right information into an assistant. Peter has a lot of properties. So he build a

8:38 dashboard for Airbnb business that tracks monthly performance stats, automatically updates and also forecast future bookings and earnings. This is a big job, right? I really love that Peter

8:49 is honest that it was not easy. I had to learn again that almost anything is possible when you approach it with a plan and determination. Two weeks ago, I would not have believed I could

9:00 accomplish such a thing. But today, we live tested and manually cross-checked the data to ensure reliability. And it is up and running. Again, manually crosschecked data. It's not the

9:13 technology, it's how you approach the technology. Ooh, Lisa. Lisa's week one commitment. Autopop populate a list of qualified leads for my husband's construction business into Posgress

9:24 database and draft not send. again responsible use a personalized outreach email to each and then of course we jump in and make suggestions how to achieve this best way. Ben is also in finance

9:38 and accounting. So, keep that in mind. This is his job, right? He can spot issues and mistakes. I drop two CSV files, a client's bank statement and their QuickBooks register, and it

9:49 matches every transaction. It can gives me only the exceptions to clear, tells me if the account ties out. It's read only. I I don't know if you see the pattern but the way you approach things

10:02 my members they literally like not send read only cross check data this is like a pattern across the people who actually use AI in the right way I still review the work and commit the reconciliation

10:15 in QuickBooks myself the part that took the most work wasn't building it was proving I could trust it Kevin landed huge deal and we helped him with you know how to structure it and just being

10:28 part of his whole journey with us. So Kevin's project is AI intelligence machine for global specialty pharma company a real decision support not decision-m decision support system and

10:40 the reason is decision support system because we have proprietary data access so they can take a lot of data and manually people would process that data by themselves which takes a lot of time

10:50 which of course AI is a really great data processor so that it can really help crunching those millions of lines of proprietary information. So a real decision support system that surfaces

11:01 anal analyzes opportunities the way the best people do. The part I love understand the human first then build the machine around them. Carbon brain plus silicon brain as peers. Amen. Okay.

11:14 Jen her first client was again bookkeeping. But what is very interesting her meeting with a client she recorded it. She plugged the audio into Claude and we came up with a better

11:26 plan to improve her current bookkeeping tools and build banking summary page that gives her exactly what to put into her QuickBooks instead of having to look several places. A client owns a very

11:36 successful massage business, but her bookkeeping is backlogged more than four months and has been notorious so far for decades. She also had no clue what her rebook rates were for returning clients.

11:49 Very real business problem. So what did Jen do? She built sales calculator, rebook calculator, deposit builder, and tip finder. When she runs her mindbody reports, she can copy paste them into

12:03 the new Google sheet and they generate data for each of those tools that Jen built. That data easily translates to her QuickBooks and several months of backlogged bookkeeping are now up to

12:15 date. Mind-blowing. Cameron is in finance as well. Cameron set up Gobot and the big piece of Gobot is a critic mode and a board of directors. He built a legal advisor to trash NDAs from my

12:28 work. When we do engagements, there are NDAs to sign and execute. Frequently there is a mismatch in clauses. The legal advisor captures the two documents, generates difference and

12:40 proposes language to close the gap. We then send to the real lawyers for review and approval. This saves time, quicker turnaround, and lawyer expense. What surprised Cameron, the critic is the

12:52 most valuable agent as it pointed out gaps in the results from other agents. So, critic mode challenge any AI output and get AI to challenge and crosscheck last before it arrives to you. Never

13:07 ever take AI's first output or any agents output at face value, especially when it's legal. Ken, my goat setup now handles my time sheets almost autonomously. So Ken is in freelance.

13:19 Yeah, he's a freelance consultant and have kept track of customers time worked for invoicing. So I keep time sheets. The old way of doing this was with apps or tools from vendors, customers, but

13:32 mostly manual work. Gobot has read access. Read access means it can't adjust. It just extracts data. data processing to my work rellated calendars, multiple calendars and can

13:43 get around 80% of the time sheet info from there. It daily checks the calendars and inserts info in the time sheet database for that day. Love the voice option via Telegram. I can do that

13:55 anywhere and it's fast and convenient. Again, we have Ben and this one is very fascinating. A goal was simple. Hand off a task I have done manually for years, usually forget to do and still consider

14:09 important. I regularly pull production data for the companies I invested. The raw export is messy and I have always cleaned it by hand before turning it into usable tracking sheet. By the end

14:20 of week one, clot code was not just completing the task. It had built a reusable script that cleans the raw file, checks its own math, validates the total against independent sources, stops

14:32 if the data looks wrong, and ask before overwriting anything. And the best finding, it flagged that 2 years of New Mexico data appeared to be mislabeled and swapped. I would not have caught

14:46 that manually. It is a second set of eyes reviewing data that matters to my work. I used to clean Texas oil and gas production records by at least three or four times a month. Now I run a single

14:58 PowerShell command and the work is done in seconds. Sean owns grooms wear shop. So physical location analyzes social context for coming trends in wedding suits reference flores venue decorators

15:11 and bridesmaid dresses for color themes as well as what pain points grooms are reporting through the planning process. So it analyzes the social media of my four to five local competitors as well

15:23 as four or five aspirational brands presenting the information in a dashboard. Push both of those so both social media analysis and competitive analysis to a Google Drive folder shared

15:35 with my social media manager copyriter. So now instead of wasting social media managers and copywriters time finding just sitting on social media or just browsing competitors all that work is

15:48 done. So now social media manager can do what they do well right to create social media content and then copyriter actually writes a piece based on actual pain points that already identified for

16:01 them. So this happens monthly. Then Shan also built a health uh agent to track weight, sleep, and blood pressure. Shan followed with a win. So my business is a suit shop and our main supplier is

16:13 German manufacturer. That already sounds like trouble. They use fabric codes and sizing which varies depending on the fabric and garment. About 18 months ago, I built a conversion sheet that uses

16:26 functions to translate the sizes. Today, I fed my conversion into Henry with his goot. Every member gives a name to their own personal assistant. So I build gobot and I gave it to my members. They give

16:39 it their own personalized name. I gave him an invoice and connected him to my ordering spreadsheet. And he not only did the conversions but told me who we had ordered the items for and if they

16:53 were purchased or to be added to our higher stock. This is a huge win for me. freeing up a lot of time and making a task that anyone can do. Okay, Carlos first paying client. Carlos is building

17:06 order system that will take all PDF orders that are sent to Gmail box, sorting them and leaving them ready to enter their order system will evolve to do this automatically. They are selling

17:18 close to 4 million and will save CO2 hours daily. Can you calculate? Well, I think we tried to help um with how Carlos should charge, but two hours for CEO daily is a insane huge saving.

17:34 Another one from Ben. So, this is in finance valuation advisor into the bot this week. A back off napkin oil and gas asset valuation engine reachable with just slash command on his phone. I

17:45 pointed it at a real closed deal and instead of guessing a number, it led with the one question that changes the whole model. Working interest versus royalty interest. Then handed back a

17:56 ranked list of exactly the inputs it needed and flagged two mistakes I had made. Interesting reflection from uh Ben is a system that produces analysis faster than its owner can act on. It

18:10 doesn't create leverage. It creates a backlog with better pros. What surprised my assistant argues with me. This is what I like the most. My go argues with me all the time. It refused to value a

18:22 deal of a missing input and then the board review the bot itself. The critic caught the other advisors calling a fix done that wasn't done. It pushed back instead of flattering. Didn't expect

18:34 that. Won't give it up. Dale set up goat and also with board meeting and critic. And at the bottom of a meeting update was a reminder. I had a hard stop at 3:30 p.m. which is when I need to get my

18:47 daughter from school. Gobot realized that the schedule was different on Wednesdays. This is what I wanted in personal assistant. And you can see here that Dale was talking with his personal

18:57 assistant completely about different thing like a project and merging. And at the bottom it said like hey by the way you have this many hours till you have to pick up a daughter. These proactive

19:07 nudges from AI that actually not pollute your brain and mind but put your attention on the right things at the right time is really hard to build but once you build it and once you

19:19 experience it it's really hard to go back to anything uh or chatbot I don't use chat GPD so these are the real stories from real people with real problems not invented problems these

19:31 people are not trying to get famous on social media or just pump a content nonstop up. They're just trying to use this technology to solve real problems that save them time so that they can

19:43 focus on other areas that are much more important than crunching data and copy [music] pasting. Okay, see you in my next