This weekend should include a few time slots to produce some epic prose. I’m still working to round out my 104 week writing project into machine learning. As that project comes to a close the next big journey will start. My five year writing plan and research trajectory will make a jump to the right and the plan will keep on moving along. One of the things that I had become very concerned about was sticking to the plan and producing a large enough degree of output. For example, just today I started to wonder if I should write a quick book called, “Your bot, my bot, our bot: A chat about platforms and bots.” Something about the latest large language models and the very real threat of content flooding has gripped my attention. Good writing comes from the depths of passion around a subject and maybe I should just give in and spend a couple of days focused on that bot project.
Central to that intellectual question is what happens when we are not able to reasonably tell if our interactions are occurring with a bot compared to a person. We may very well be able to set up a friend bot and text with it all day shortly and it will be hard to tell if it’s not your college friend or long lost work associate. People are already trying to use a collection of video responses to generate virtual versions of a person. Enough video of me exists online that I’ll probably end up being a video bot one day. Maybe my current focus is about what will happen with that next persona and my concerns if it will end up being accurately polite and sardonic. Most of the bigger bots that have been created over the years and launched by even large companies have taken a turn to very mean rhetoric.
We all know that a lot of internet forums and other online exchanges are not the best places to find civility. Training anything from that type of content is going to be a mirror to it and not as much a leap forward toward a vision of a civil society where a shining city on a hill remains delightfully just in reach. My corpus of writing based on previous GPT models ends up producing thoughts about writing, the process of writing, and complaining about both writing and the process of writing. While I tend to write about those things on this weblog and that is why that content exists in the corpus, my discussions with actual people during live dialogues generally don’t go that direction. People don’t really want to hear about the troubles and tribulations of the writing experience. That type of nonsense is best left for the written page where it can be ignored online or pursued if necessary.
Within those central questions you can probably sense that my focus on a storm of modernity and bots could very well be a chapter in a book about the intersection of technology and modernity. My guess at the moment is that it could also be a stand alone manuscript and might be interesting. Either way it is probably an area of focus that will end up on my updated writing plan pending the completion of my current machine learning book. It feels oddly cliche to have collected enough content to publish a book on machine learning. A lot of those types of books have come into being in the last couple of years. I don’t think any of them are written in the same style or cover the breadth of content that I have evaluated, but that does not negate the sheer volume of machine learning content that has sprung up into being recently.
This very weblog contains the two parts of my five year writing plan. First, it includes a reasonable list of upcoming research which describes the backlog of planned things. Second, a page on the weblog is devoted to my research trajectory. That collection of thoughts has been pulled together to help describe the general content areas that are pulling my attention from time to time and end up informing future publications. Together those two things are what fuel a five year writing plan which is really a way of measuring my writing output against both the possible and the rate of my actual production. Overall this is an important way to hold myself accountable to the possible creation of content. It also creates a process where I’m managing my time and screening out things that should not receive my time and attention as they don’t contribute to building something meaningful. You are probably well aware that my goal is to work just beyond the edge of what is possible. It’s just on the other side of that edge of possibility that the remarkable awaits.
Today I finished working on the main content for week 88 of The Lindahl Letter. That one is a bridge piece between two sets of more academic side efforts. I went from working on introductory syllabus to starting to prepare a bit for the more advanced set of content. Initially, I had considered making the advanced versions a collection of research notes that were built around very specific and focused topics. That is entirely a path that might be taken after the 104th week. The packaging on the content instead will be put into another companion syllabus to allow an introductory look and a more advanced topic follow up for people looking for a bit more machine learning content. Functionally those two documents put together will be the summation of 104 weeks of my efforts in the machine learning space. It is the book end to my journey into really diving deep into machine learning and studying it every weekend and a lot of weekdays.
After two years of digging into the machine learning space I’m going to pivot over and focus on writing and studying artificial intelligence in general for year 3 of The Lindahl Letter. It should be a fun departure and hopefully it will mix things up a little bit with a broader collection of literature. A lot of people talk about deploying AI in the business world and almost all of that conjecture is entirely based on deploying a machine learning model into a production environment. When those same people deploy an actual AI product they will hopefully see the difference.
Yesterday I spent some time thinking about my research trajectory and where my writing efforts should be placed. Intellectually I know that I should use the conference cycle to help motivate my writing efforts and keep them on a tight publishing schedule. Over the last few years that has not really happened in a consistent way. I have written and put together content for a series of talks, but I never finished taking that content and putting it into a journal article format or working it into a conference paper. That is probably one of the first things that should be on my writing plan that is in the process of being reworked. My writing schedule works and I turn out content every week. That is a proven effort at this point. It has worked for well over a year within the machine learning content space. All of that content exists in a well contained Google Doc with weeks 1 to 104 planned out for Saturday delivery. At the start of the year, I did take the first year of that content and put it into a manuscript form and it was edited by a professional to make sure it was ready to be shared in print. That cycle will continue until the end of the 104 planned posts.
Please keep in mind that my writing plan is not a theme of the year or anything like that. It should be an organized and thoughtful research trajectory from the start to the finish. The other constraint that I put on it is to very clearly view it as a measure of what I could do with the time that I have in front of me. If I only had 5 years to muster up writing efforts, then what should that time be spent on and the most important things should be closer to the beginning of the journey than the end of it. Time is incredibly unforgiving and before you know it from our perspective things will move along with or without the time being put in at the keyboard to create resplendent prose. Within the moment we know time is about to pass, but we have the ability to only elect action or inaction.
My 5 year writing plan as of March 3, 2022:
- Year 1 – Heavy ML focus for the rest of 2022
- Finish writing a collected series of ML/AI essays on Substack and combine them into a manuscript, “The Lindahl Letter: On Machine Learning Year Two.” This manuscript should include both years one and two.
- Weekly Substack posts
- Manuscript generation at the end of the year
- Will need to be edited by a professional before the print edition goes live
- Rework last years speaking engagement talks into academic papers. This could be one combined paper or potentially 5 different papers depending on how the initial effort shapes up.
- “What is ML Scale? The Where and the When of ML Usage.”
- “The ML scale problem: Thinking about where and when to use ML, ROI models, synthetic data, repeatable frameworks, and teams.”
- “Applied ML ROI – Understanding ML ROI from different approaches at scale.”
- “Demystifying Applied ML – Building Frameworks & Teams to Operationalize ML at Scale.”
- “Figuring out applied ML: Building frameworks and teams to operationalize ML at scale. V3”
- Rerun the MLOps Github research and turn that content into a paper
- Year 2 – For 2023 I want to pivot into studying sentiment analysis and modern polling methodologies. At this point, I will have written 104 essays on ML/AI and should probably refocus on a specific topic that is material to ML/AI, but adjacent to it as an area of research. It’s possible by 2023 that quantum computing will be a huge topic for research and will end up getting some attention as well.
- Automated sentiment analysis paper
- Sentiment analysis and machine learning essays for Substack
- Modern polling methods essays for Substack
- The breakdown of modern polling paper
- Year 3 – 2024 will include a return to writing about local government administration and technology. It will be 20 years since earning my master of public administration degree. By this time my writing should be as crisp and focused as it will ever be and my perspective on technology will be well considered from my previous work on ML/AI.
- Technology and local government administration
- The intersection of public administration and technology
- How technology influences the practice of governing
- How government uses ML/AI technology
- Year 4 – 2025 will probably be the year where quantum computing has broken down modern encryption frameworks.
- Changes and uses in encryption technology
- Encryption and society
- Quantum encryption
- Year 5 – 2026 is going to be a year where my backlog should be highly full. The previous 4 years of this writing plan should have created a ton of leftover writing works.
- A reflective work on ML/AL
- Did open source MLOps technology survive?
- Did the serverless trend pan out in the cloud?
Maybe now is the time to have a whiteboard session and rework my planned academic article list. This could be as simple as making a list on a blank page of paper or it could get a lot more complex. I’m thinking taking the complex road might be the way to go with this one. To that end, right now I’m wondering about the top 5 articles I would like to sit down and write. I can begin to see a few of them in terms of structure and breakdown, but none of that inspired me to stop working on this post and begin the process of writing. I’m going to need to focus on two directional elements. First, what is my academic writing trajectory and what are the results in terms of article output that would arise from that path. Second, what are the articles that I would really like to read that would be groundbreaking in some way. Plotting out both batches of content should help me find somewhere where either some overlap occurs or maybe somewhere where a little bit of intellectual stretching could get me closer to the edge of what is possible and avoid derivative muddling.
In order to start that effort I’m going to map out my general areas of research interest. The last time I did that in a serious way was back on September 5, 2021. I’ll take that base and begin to expand it to the next level. You could build your own base by quickly making a list of the top 5 research interests you have. From those general interests you will find that you want to add more words to the topic to shape it into a more specific and targeted point within that larger topic.
- Public administration
- Local government administration
- The intersection of public administration and technology
- How technology influences government
- How government uses technology
- Changes and uses in encryption technology
- Encryption and society
- Quantum encryption
- Sentiment analysis and modern polling methodologies
- Automated sentiment analysis
- Sentiment analysis and machine learning
- Modern polling methods
- The breakdown of polling
- General uses cases for machine learning
- Common API use cases within the ML space
- General ML use cases compared
- My general look at MLOps open source code
- A review of MLOps Github repos
- The ethical use of large language or foundational models
- Language models and society
- The intersection of technology and modernity
- Oversupply of information (flooding)
Those areas of research interest have a trajectory. Each of those general themes is going somewhere or evolving somehow. To capture that I started to draw out the 5 topics above and consider what’s next. What would be the bubble next to these 5 bubbles. How do those bubbles interact and in what general trajectory are they starting to move? Building out that series of relationships helps me think about what areas need the most consideration and where the most movement is about to happen. I’m going to spend more time today on that mapping and whiteboard effort. It is not something I can do with a pen and paper as a lot of give and take needs to occur within the live editing consideration brought to that type of landscape.
Ultimately, understanding that general research trajectory is key to being able to complete the action described above that inherently stretches things toward the possible and away from the derivative.
Reworking all of that made me wonder if I should just pull up my 5 year writing plan and see if it still makes sense. At some point, I’m going to need to rewrite my 5 year writing plan and figure out what things on that list really deserve my time and attention.
Apparently, I had forgotten about making a new static page on the weblog devoted to upcoming research. It already contained over 10 items on the list of work I’m supposed to be completing. Right now I’m looking at several different things that are lined up about what I’m supposed to be working on and they are all somewhat interesting.
- A research trajectory summary
- A writing schedule plan
- A list of upcoming research (without any prioritization)
Right now all 3 of those things have been made into static pages on the weblog. The most straightforward part of my planning trifecta (research trajectory statement, writing schedule, and upcoming research plan) of thinking about what I’m going to do next is really the writing schedule. It really just details my plan each week to sit down and be productive at the keyboard. For better or worse that means tracking in advance what my weekend mornings are dedicated to working on and how that time will be best spent. My writing schedule can be summed up as a simple look at weekends vs. weekdays and what needs attention.
The upcoming list of research ideas is really just a pile of problems for future consideration. It represents for better or worse a parking place for ideas in need of more attention. That means at some point on some weekend they are going to get the attention they deserve or maybe they will just be abandoned in favor of something else. I only have so much time and attention to spend on things and some items are going to get more of it than others.
What I am going to spend some time on today after reviewing the Substack post that was written yesterday for grammar and clarity will be to revise my research trajectory statement and try to get it posted online. I think that is really where I need to spend my focus for the day. It might very well involve a little bit of time with the whiteboard and a little bit of time writing up my efforts after that exercise is complete.