A Change of Direction for Lukich.io
Hello to everyone, and welcome back for another action-packed post.
I hope that you’re all off to a great start to the year. It’s May. The weather is getting warmer. The grass is greener. And I’m trying to spend as much time outside with the kids as possible.
It has been an eventful first 4 months of the year for me — both professionally and personally. I have a lot to share. Most notably, I’m going to be making a significant directional change to this website. There is a lot that went into this decision, so in order to provide the necessary context, I’m going to have to back up to January.
Poker: January to March
When I wrote my last post in January, I was focused on improving my MTT game. I had released a new module for Solver School, From Macro to Micro, which focused on building a data-driven strategy, and I was using that framework within my own studies.
In Michigan, the weather gets cold during the first few months of the year. The days are short. There are far fewer events and family outings. Much more time is spent inside at home. As a result, the days can somewhat blend together.
While tedious, I find that can be the right mix of environmental factors to get a lot accomplished. It’s much easier to keep your head down and grind when nothing else is going on.
And grind I did. I spent January and February working on getting better at single-raised and 3-bet pots in the 20-50 BB stack size range. I also spent a few hours per week consuming training videos. I purchased Patrick Leonard’s outstanding MTT video course, Pads on Pads, and was using that as a complement to my own work. Finally, I was playing in a lot of online tournaments on the Michigan-facing sites — PokerStars, WSOP, and BetMGM.
As a quick side note, my playing mix has changed significantly over the past few years. I was primarily a live cash player until 2020, but COVID pushed me online. Even though things have opened up since 2021, I haven’t been able to fully reintegrate live poker into my regular schedule the way it was pre-COVID when I lived in Maryland. While I occasionally make it to the Detroit casinos for a live cash session, most of my current volume is online.
I kept a busy schedule of playing MTTs 2-3 nights per week during January and February — heavy volume for me. Unfortunately, the good results didn’t follow. Instead, the downswing that started during the second half of 2022 continued.
I started playing tournaments in 2020 and had some really good success over my first few years, with several 5-figure online scores. But since the fall of last year, I’ve been slowly losing. I’m cashing tournaments at similar frequencies, but the final tables and wins haven’t been there. With the top-heavy nature of prize pools, the lack of podium finishes has been the difference between winning and losing.
This is nothing unexpected. Tournament poker is somewhat insane, and the variance can be significant. I remember reading this analysis on 2+2 back in the day. This article from 12 years ago quantified the crazy variance in playing tournaments. Using results from Shaun Deeb from 2011, the author of the piece simulated 1,000 Shauns each playing in 1,000 tournaments. He found that even someone at the skill level of Shaun Deeb would have a losing record over 1,000 tournaments ~13% of the time.
I’m certainly not as good of a tournament player as Shaun Deeb (now or 2011 Deeb), so if someone of his skill level can hit an extended downswing through the course of normal variance, I obviously can too.
Additionally, my volume is minuscule compared to the scale referred to within the article. 20 MTTs is a heavy week with my schedule. So a downswing can last months of actual time. Losing over an extended period of time really takes a toll on my confidence and ability to stay positive.
In March, I went to Vegas on a yearly trip I have with some friends to watch the 1st round of the NCAA basketball tournament. It’s the trip I look forward to most every year. I get to hang out with friends I don’t get to see as much as I did before we had families, watch and gamble on basketball, and play a lot of poker.
If anything could get me out of my funk, it was a bunch of live cash poker. After all, this was my bread and butter. I’ve been playing mostly online for the past few years with mixed success. But my comfort zone is playing $5/$10 NL cash games in a casino.
The trip started off great. I had three killer 6-hour sessions over the first three days. But a disastrous 4th day followed by several unsuccessful bullets fired into a tournament turned a few days of wins into a frustrating breakeven trip.
One thing I know about myself — I’m not wired in a way to mentally handle the grind of full-time poker. I struggle not to let the emotion that comes with losing impact the rest of my life. If I were playing for a living, I question whether my mental health would be able to take it. Even though I’m currently only playing ~5-10 hours per week, the lack of positive results for such a long period of time was starting to take its toll. I was getting burned out.
I’ve always prioritized being a good dad and husband. That’s my first and most important job. I also work a full-time job as a marketing data science consultant, which keeps me busy during the day. It’s hard to find much free time in my schedule. Most of that time I did have was spent studying, building out the products I sell at Solver School, and playing poker. Without some positive results built into feedback loops, I struggled to maintain a positive attitude.
Taking a break
I mentioned that I had fired 2 bullets into a tournament. On the last day of my Vegas trip, I played one at the Wynn. Four of the guys I went there with had left to fly home that morning. Only one other friend, Kevin, was sticking around with me until Monday. Kevin had made plans to see his sister during the day, so I decided to spend my day playing in the tournament.
I never got much higher than the starting stack on either bullet. When I busted for the 2nd time around 5 pm, I still had about an hour left of late reg — I’d still have around 20 big blinds to work with if I wanted to enter a 3rd time. But my head and heart weren’t in it, so I decided to quit for the day. Kevin was getting back to our hotel at 8, at which point we would go out for a fun night of dinner and drinks to close out the trip.
I decided I’d enjoy the downtime before dinner. I went back to my hotel room and opened up my iPad to write in my journal. I built a section within my Notion workspace to journal in over the past year. I’ve been doing it semi-frequently and have found it helpful to gain clarity in some of my thoughts. As I emptied my thoughts into writing, I acknowledged to myself that I needed to take some time off from the game.
After clearing my head, I laid down on my hotel bed and scrolled through Twitter. OpenAI’s new AI large language model, GPT-4, had just been released, and it was dominating my feed.
I had nothing to do for a few hours, and I’m a sucker for cool technology. So I read a thread. And then another. And before you know it, it was 8 pm. I had spent a few hours reading all about the amazing things ChatGPT could do.
I read about this guy using GPT-4 helping this guy to bootstrap a business.
I read about how it scored in the 90th percentile on the Bar exam.
And another guy who used it to build Pong in 60 seconds.
I tried the GPT-3 model in December and was impressed. But this felt way more impactful. OpenAI’s language model was now doing extremely difficult and complex tasks with ease. It was week 1 and already, the use cases seemed somewhat endless.
While I’m sure I could have spent all night scrolling Twitter, it was also my last night in Vegas. So I shut it down for the day to enjoy the evening with my friend. We ate some good food, drank a few too many beers, and had a blast splashing around at a $1/$3 table late into the night.
On my flight home the next day, I couldn’t shake GPT-4 from my mind.
There was a rush of excitement racing through me. I couldn’t remember the last time I felt this giddy about a technology. As I continued to scroll through Twitter on my flight home, my mind began to race through all the potential things this tool could help me to do. I wanted…no, I needed to learn more.
I now had my answer as to how I would fill my time away from poker. In the last hour of the flight, I opened up my journal for another entry. I decided to take the entire month of April off from poker and use it to learn more about AI. I wanted to understand what it was, how it worked, and most importantly, how I could use it to improve different areas of my life.
AI research
I didn’t play a single hand of poker in April. I didn’t review any training content. I didn’t watch any live streams. I didn’t plan my Sunday evenings around the PokerStars MTT schedule. I didn’t run any solves. I didn’t even open a solver!
The majority of my free time outside of work and not with family was spent diving into AI. Looking back, it was exactly what I needed. As I dug deeper into the rabbit hole, I did some awesome things.
First, I built a web app.
Back when I was developing the From Macro to Micro module for Solver School, I came up with an idea for a web app. Within the course, I was using PioSolver’s aggregated report functionality to look at data on a larger scale. The problem is that the PioSolver’s default reporting output is not all that useful for analysis. Without additional metadata to help describe the characteristics of each board, it becomes challenging to do any large-scale data analysis.
I have maintained a lookup table that I’d use to append to these reporting outputs, but the process of joining data and building out the same views was repetitive and time-consuming. I had an idea at the time to write an R script that would automate this process.
I started experimenting in January, but the project was going slowly because I’m not that great at writing code.
Enter GPT-4. One of the amazing things it could do was take input prompts and turn them into working code. After spending a weekend working with GPT-4, I made so much more progress than I had in several months of stumbling on my own.
I designed and coded a working Shiny app from scratch that accepted a PioSolver aggregated report as an input and generated an enriched data table with a bunch of metadata attached to make analysis easier. I even wrapped it up in a nice UI and built out a couple of useful data visualizations. You can try it out for yourself here. It’s not a professional app, but for a simple solution that took me a few hours on a weekend, it works pretty well!
Then I wrote an eBook.
I began to see so many ways in which I could use this in my day job as a data scientist.
Build a project plan? Check.
Write some SQL code dealing with unique variable types? Yep.
Outlining slides for a PowerPoint presentation? Good too.
I started thinking about all the incredible ways in which all business professionals could use this in their day-to-day work. It felt like I had an incredibly smart intern at my fingertips who happened to know a lot of things about everything and was eager to help me get things done.
As I experimented, I learned more about the nuances of writing prompts. I experimented with a bunch of different prompting styles to figure out the best way to interact with the model. Through trial and error, I think I was coming up with some excellent prompting techniques to help my own work.
This was a matter of education. If I was facing these challenges, others surely were too. I decided to share some of these thoughts in an essay about using GPT-4 within marketing. That extrapolated out into other areas of business. The next thing I knew, I had an outline for an eBook.
After some iteration over a couple of nights, I used GPT-4 to outline and write an 80-page eBook aimed at helping people across all functions within enterprise organizations use the technology to help them as well.
The big caveat here — while GPT-4 is great at generating outlines and first drafts, it’s not equipped to write a full 80-page eBook autonomously. At least not yet. But it’s excellent at getting the ball rolling.
What would normally take weeks or months of writing was done in 2 days. I still had to edit and format, but I significantly reduced the time to get this project completed.
The title of the eBook is The ChatGPT MBA: Using ChatGPT to Maximize Productivity in Your Business. Here’s the cover page and the table of contents:
I am putting the finishing touches on it now and plan to release it in the next 1-2 weeks. I haven’t decided if I’ll sell for $5 or give it away for free to drive email addresses. Possibly a combination of both but more to come soon!
And currently, I’m building an app to chat with data that sits in a back-end database. I’ve been diving into the Langchain documentation and learning how to connect agents with LLMs and other tools to perform more complex tasks. It’s a work in progress, but something that I never thought possible before — particularly for someone like me who can’t code well.
In addition to the projects above, there were a number of smaller ways in which I integrated these solutions into my day-to-day life.
My wife and I are planning a trip for our 10-year wedding anniversary, so I used it to plan a 1-week trip to France in the fall
I asked it to explain countless things to me — concepts, code, it is great at distilling complex topics into simple language
I summarized lengthy text — GPT-4 is great for working with a lot of text and synthesizing it into smaller bite-sized chunks
I used it to brainstorm business ideas, topics to write about, tweet threads, and much more
It started becoming a staple in my day-to-day processes.
A Return to the Felt and An epiphany
At the end of every month, I generally spend about 30 mins reflecting in my journal about the previous month and plan out what I want to accomplish in the next one. When I reflected on April a couple of weeks ago, I realized that I hadn’t missed poker nearly as much as I thought I would have.
That surprised me. After all, poker has been a central part of my life since 2018ish when I first opened a solver.
Upon further reflection, I came to the realization that my current decision to cut back on poker was not unprecedented. My first obsession with the game was from 2002 to around 2007. During that time, the Moneymaker boom had swept me (along with many others) up in its excitement. However, towards the end of that period, I became burnt out and other priorities took precedence in my life. Consequently, poker transitioned from being a primary fixture to a serious hobby.
Over the following decade, my passion for the game never dwindled. Although I played more casually, I continued to engage with it regularly and delved into literature to explore new concepts and strategies.
I’ve come to the realization that I will always love this game. I want it to be a part of my life, and I can’t imagine a world in which I give up playing or learning about it. But at the moment, I don’t want poker to be my central focus.
I will continue to play. I’ll contribute to the community. I’ll work on advancing the thinking I’ve released through Solver School. But it won’t be my first priority. I won’t devote as much time to it as I had been previously. And just as important for my own mental sanity, I won’t put any pressure on myself to meet deadlines. Poker is fun to me — it’s something I love. But as soon as it turns into a responsibility, the love starts to fade away. It becomes work. And that’s the point at which I know it’s time for me to take a step back.
How Poker fits Into My Future Plans
So what next? Well for one, I’m going to continue playing poker.
When May rolled around, I started playing again. I’ve since settled into a schedule of playing 1-2 nights of online MTTs per week. I also made it to the Detroit casinos for a live session a week ago.
The sample size is quite small, but my time off has cleared my head nicely and has translated into some good success in my first few nights back. The positive outcomes were highlighted by a deep run in a $500K guaranteed event on PokerStars a couple of weeks ago. I’m keeping my fingers crossed that it’s a sign of breaking out of the downswing.
I will also continue to contribute to Solver School. Ever since launching last January, Solver School has turned into a nice side business. There’s an awesome community of poker players and data nerds like me who have signed up. I’ve also built out a consistent group of coaching clients who want to better use solvers in their own studies. While I don’t have any huge product releases planned, I do plan to create additional modules from time to time as an addition to the library of content.
Finally, the updates bring me to this blog. When I first started this website in January 2020, I envisioned it as a space to explore my data-driven poker work. More specifically though, I wanted to explore my passion. And exploring the game of poker through the lens of data was my passion at that time.
Over time, people change and passions shift. While I will always love poker, I can’t say that it’s the thing I’m most interested in at the moment.
As a result, I’ve decided to shift my focus to keep this website aligned with my current interests. And at this present time, I’m all-in on understanding AI, how individuals can use it to drive productivity gains, and its impact on organizations and society. I have a lot of thoughts I’ve been jotting down, and I plan to use this space to explore and share them more broadly.
Moving the Poker Posts from Lukich.io to Solver School
You’ll notice some changes on this site in the coming weeks. Specifically, I plan to shift its focus to AI. This is obviously a space that many are rushing into, so I’ve been thinking through this carefully to make sure that I approach it in a way in which I can add value. I’m less interested in the specific things that AI can do. There are plenty of Twitter threads and newsletters that detail all of the amazing capabilities that these tools have.
Instead, I’m more interested in the broader implications of AI — specifically, what it enables us to do and the ripple effect through society. I want to explore the impact of AI, both for individuals and organizations, and theorize as to what that means for business, markets, and society.
In the upcoming weeks, I plan to temporarily shut down this website while I rebrand and build it into a streamlined place to host my AI thoughts. I plan to write some long-form content about 1 time per week. I hesitate to call it a newsletter just yet, but that’s at least where I’m heading.
I also don’t want to lose the content on this website. I’m really proud of my writing over the past few years, particularly the poker-centric content that continues to drive new traffic to the website.
I plan to repost all poker-related articles on the Solver School website, consolidating all of my poker work into one place. Maintaining both sites has been a bit challenging over the past year. Despite the two different sites, there was obviously a significant overlap in content. So this is a way to bring everything under one umbrella. I hope that the traffic that makes its way here to learn about things like formation analysis will now simply go to the Solver School site instead. I probably should have done that in the first place, as I’m sure that would drive more sales of my course!
I’m also going to move the free products on this website to the Solver School domain. They’ll continue to be free to anyone who signs up for my mailing list.
Finally, I plan to retire the Excel workbook I sell on this site. So when I shut this site down, that product will be taken offline for good.
KEY TRANSITION Dates
I plan to launch the rebranded site within the next month. I want to make sure I effectively move content to the Solver School domain. So that you’re not thrown off by timing, here are some key dates:
Week of May 15 - Content moves to Solver School
Week of May 22 - Lukich.io goes offline
Early June - Lukich.io rebranded to focus on personal projects (AI, data, poker analysis, etc.)
I’m combining mailing list platforms as well, so if you want to make sure you’re getting my Solver School posts, make sure you sign up for the mailing list there!
Conclusion
I think that just about does it for my updates. I know…a lot going on. I appreciate you sticking with me to the end here. I’m excited about this new challenge I’m undertaking.
Please follow me on Twitter or sign up for the email list at Solver School to get the heads up anytime I post there.
If you have any questions, comments, or want to discuss anything, please feel free to reach out to me on any of those channels or email me. Thanks for reading.
-Lukich