I hope all 9536 of you had a great week 🔥
Welcome to the 25 new members of the TOC community.
If you're enjoying my writing, please share Terminally Onchain with your friends in Crypto!
Happy Friday everyone!
I went through a major work crisis last week.
To a certain extent, I'm sure many of you have gone through this train of thought as well.
It goes something like this.
For a few months you're locked in and killing it at work. The metrics are growing. Your confidence is increasing. And you're feeling just the right amount of learning obstacles to overcome and grow.
Things start feeling comfortable and you're in a weekly routine.
But soon the thought creeps in that maybe you're....a bit too comfortable?
You start seeing others on the timeline flexing their updates and you can't help but compare your work to the flashiness of their announcement.
And before you know it, the dominoes are starting to fall.
Have I been wasting the last few months? What have I actually achieved? Is this even the right vertical to be in? Ugh have I been spending way too much time on Twitter?
Here's the thing though. As with any crisis, it sucks in the moment. But once you walk out the other side, it actually feels damn refreshing and there's a new boost of energy you have for a couple of months till the cycle repeats.
So the good part is that I have a lot more clarity on a few things and I'm excited to share them with you all. Maybe some of the thoughts will help you as well!
None of these insights are groundbreaking to be clear. The more important thing is that they clicked for me in a serious way.
I feel a true sense of urgency to improve on all three of the things below.
Sections below:
Update my inputs (deep reading)
Learn smarter (deep research)
Evoke emotion (deeply personal)
Let's dive in.
Update my inputs
For anyone who has been following my journey the last year or so, I've had a pretty consistent pattern with my writing. I would cover anything interesting that was happening in the crypto space. Stablecoins, zkTLS, EIPs, Farcaster, prediction markets, etc.
My posts were not for beginners but they also weren't the ultimate primer that got excessively deep. Let's call it 80th percentile coverage.
I do posts every Tuesday and Friday. The flow is: scroll X, pick topic, read/cram necessary material, write up the post, and repeat every three days.
And I'm glad I did that for my first 100 posts. I knew my writing wasn't amazing. And I knew my insights weren't anywhere close to Ben Thompson level. And that's okay. The whole point was to build the habit of publishing, explore a variety of topics, understand my research framework, etc.
Mr Beast tells aspiring YouTubers to make 100 videos and he'll give them feedback and advice.
2 things happen.
98% never get close and give up.
The 2% who do, no longer need his help.
What's the lesson here?
2 things happen.
98% never get close and give up.
The 2% who do, no longer need his help.
What's the lesson here?
3,319
I remember seeing this quote and locked in on getting my 100 posts out. And it worked!
There's a clear niche, mission, and goal I have now.
Many consistent TOC readers have seen me get totally nerdsniped by Decentralized AI. It started with the agent craziness in September and since then it's been a rabbithole that has fully hooked me.
I'll repeat my goal that I mentioned in a previous post:
My mission in these upcoming months is to become the crypto x AI guy.
I want to be the subject matter expert that glues both of these sectors together by understanding what's happening at the frontier of open source AI and decentralized AI.
In order for this to happen, I need to be reading a lot more research papers and books. That's the only way I can become a true 1% subject matter expert on DeAI.
I've "maxxed out" how much I can learn from scrolling through other people's thoughts and basic blog posts. If I need to form novel insights, it's time to update my inputs.
This means intentionally sitting in the silence and reading. And then reading some more. In fact, the reading I'm doing should hurt my brain. It should feel challenging.
I won't lie. The 30 page arxiv pdfs feel intimidating. I mean, there's clearly a reason it's so hard to become the top 1% in any field right? There's only so much you can learn trying to get by with executive summaries.
But there's clearly ways to get rid of that fear. I don't have to go down new rabbitholes the same way I did before.
Learn Smarter
I couldn't help but notice how slow my learning rate has been the past few weeks. It felt a bit pathetic. Like why is it taking me so long to get through topics x, y, and z? And that honestly felt discouraging and made me question this whole DeAI thing for a bit. Maybe I'm not the right person.
And then on a walk, I had the most obvious epiphany. Why am I not AI maxxing? Don't get me wrong, I use chatGPT just as much as the average joe on tech twitter. But if my job is to literally research, I should be doing everything I can to work smarter and expedite my learning process.
So last week I bought the $200 OpenAI deep research subscription.
There's no excuse for me to have extended timelines on learning new material. It's cope. There's a hammer available to me but I'm still banging the nail with a rock.
The goal is to learn the smart way and be fast about it.
Those who quickly adapt to the new methods of consuming and learning information will have a headstart on the way things are headed. And it's clear that even a few months of using these tools before others is enough to make the margin larger than most of us realize.
It's only been a few days and I'm already loving this new deep research experience. I'm getting through material wayyy faster. For example, in just two days I was able to go through all 3 DiLoCo papers on the docket. This would have normally taken me at least a week.
And yesterday, I prompted deep research to build me an entire primer on the open source AI side - key companies, metrics, business models, etc. I basically started with one tab and let deep research guide me through the maze.
On a recent Dwarkesh episode, Satya Nadella mentioned something that I literally can't get out of my head.
The key skill is not research anymore. Rather it's having the big picture framework in your head and being able to make quick decisions on what the most critical thing to learn next is.
You point in the right direction and let AI shine the light.
Another thing to note. If you're worried about prompting and making your money's worth. Even that's not an excuse anymore through multi-step prompting!
Bucco has a fantastic template I've been using to get started and the results have been outstanding. Over time, I'll of course slightly tune this to increasingly fit my needs.
People were curious, so here's how I'm using Deep Research. I'll walk through the prompting and then an example:
1. First, I used O1 Pro to build me a prompt for Deep Research to do Deep Research on Deep Research prompting. It read all the blogs and literature on best practices…
1. First, I used O1 Pro to build me a prompt for Deep Research to do Deep Research on Deep Research prompting. It read all the blogs and literature on best practices…
And like clockwork, just yesterday, the legend himself once again dropped the perfect video at the perfect time. I'm going to be going through this over the weekend and want to be using LLMs like Andrej. Why am I doing less than 1 prompt a day while Andrej is saying he's constantly using it in his workflow?
New 2h11m YouTube video: How I Use LLMs
This video continues my general audience series. The last one focused on how LLMs are trained, so I wanted to follow up with a more practical guide of the entire LLM ecosystem, including lots of examples of use in my own life.
Chapters…
This video continues my general audience series. The last one focused on how LLMs are trained, so I wanted to follow up with a more practical guide of the entire LLM ecosystem, including lots of examples of use in my own life.
Chapters…
Evoke Emotion
I highly recommend all of you to read this tweet by David Perell - it's incredibly well written.
This AI boom has set off an existential crisis in me.
Some background: I’ve been teaching writing for the past six years. In that time, I developed frameworks for how to write well and a reputation as a good teacher to learn from. Partially because of the AI wave, I decided to…
Some background: I’ve been teaching writing for the past six years. In that time, I developed frameworks for how to write well and a reputation as a good teacher to learn from. Partially because of the AI wave, I decided to…
Since I started actively writing in the crypto space 4 years ago, I've always considered my job to be the guy who breaks down technical concepts. Read whitepaper, translate to english, and provide some fun context around the new subject.
But it's becoming clear to me that doing "deep dives" are becoming increasingly futile.
The amount of expertise required to out-do an LLM is rising fast. For example, the quality of a well-prompted, ChatGPT Deep Research report is already higher than what I can produce in a day's worth of work on almost any subject.
It feels like a waste of time for me to break down any whitepaper considering all of you have an on demand custom tutor that can explain it any way you'd like.
I know ChatGPT has been around since the end of 2022 but it only recently has started feeling like the tools have gotten so good that they're worth it for most complicated research now (personal opinion of course).
If that's the case, then what's the moat for writers like me?
The question is: What kinds of non-fiction writing will continue to last? Here’s a heuristic: The more a piece of writing comes from personal experience, the less it’s likely to be overtaken by AI. Personal writing, like biographies and memoirs, aren’t going away anytime soon.
That's because people have data about their lives that LLMs don’t have. Having a unique perspective helps too. This is Peter Thiel's famous interview question: "What very important truth do few people agree with you on?” If you have an idiosyncratic way of looking at the world, you don't have much to worry about.
The common thread here is humanity. People are also interested in people. Their stories, their struggles, their emotions, their drama, their unique insights into how the world works.
To me, it seems like 3 things are worth focusing on and getting really good at:
Being as authentic about my learning journey as possible. Sharing all the highs, lows, frustrations, obscure epiphanies, etc.
Getting really good at asking unique questions and being able to articulate insightful opinions / takes as effectively as possible.
Curating the best takeaways from the most niche content in my vertical. Finding the gold nuggets that most people don't have time to come across.
So what does this mean in practice?
My guess is that the north star metric now goes from # of posts published / week to # of takeaways provided / week. I know takeaways is a vague term. But I would define it as providing concise, meaningful commentary on something I researched or consumed.
I imagine that some of my posts may look like Byrne Hobart's weekly longreads format or Packy's weekly dose of optimism. These kinds of articles will force me to focus on increasing the quality of my content inputs and help me train the active reading muscle. Meaning, I'm actually elaborating on ideas proposed in a blog post or providing counter-arguments to a segment of a podcast episode.
In all honesty, I'm not exactly sure what this directional change ends up looking like in the posts. BUT! What I do know for a fact is that I'm going to be treating my writing that I send out less as deep dives and more as updates on my learning journey. They may feel like the same thing but I believe the subtle difference is actually huge.
I don't just want to explain things I'm learning. I want to seriously feel passionate about what I'm doing and convey that enthusiasm through text. All of you aren't just reading a newsletter covering decentralized AI but rather joining me on an intellectual journey that resonates with you.
For example, I was hesitating to send this post out because I felt a bit cringe about sharing "my crisis" and "lessons I'm applying". In my head, the content wasn't technical so I thought what's the point?
Wrong!
If anything, my guess is that these kinds of posts will resonate the most with all of you because they're thoughts that you simply can't reproduce with ChatGPT. This write-up was a direct result of my crisis that I had the last two weeks and how I got through it.
That's it for today's post.
I hope all of you have a great weekend and touch some grass!
I'll see you all next week 🤝
- YB
love this post so much YB!
this is your post that has brought me the most value by far. 🔥
Yesterday's TOC deep dive some simple lightbulbs that went off in my head, but feeling a sense of urgency to act on them https://terminallyonchain.xyz/crisis
@colin @reidtandy odd formatting with the top cut off here on the frame. also noticed the callout sections are not showing the text correctly when there's bold or underline in them
We're aware of the top one, but thank you for the nudge! Will look into the callout section - thank you!