Discover more from Morfternight
Morfternight #66: Happy New Year!
The one about weeks, decades, AI, and rockets.
🤩 Welcome to the four new Morfternighters who joined us last week.
We love to have you here, and I hope you’ll enjoy reading Morfternight.
If you do, remember to share with a friend by clicking on this button.
📷 Photo of the week
Spinning ’round - More Photos
👋 Best Wishes from Vienna!
As I mentioned last week, I am not a massive fan of the couple weeks that just passed, which is why January 1st tends to improve my mood each year.
Not because I believe that the date has any particular significance but because it marks the end of the agitation and the return to normal.
I don’t want to discount the importance of years for astronomers or astrologists (although I discount the importance of astrology as a whole, so there’s that). However, years don’t make sense to me to move forward in life.
Use weeks to make progress.
Weeks contain enough time to do something and are at the same time short enough to immediately learn from what we did.
Tasks should take less than one week, and anything more significant should be considered a project and broken into smaller tasks. Of course, I am not only talking about work; that also applies to one’s personal life.
It’s very easy to be overwhelmed by the magnitude of larger tasks, so starting them is difficult enough to be too big of an obstacle. But, on the other hand, if one is strong enough to begin and make progress, the danger is that keeping a set direction for an extended period without a feedback loop makes a small initial error into a big mistake.
By breaking down anything you set to achieve into small enough chunks that they can be done in a week, you make it at the same time easy to start, to progress, and to regularly check if the direction is still correct.
Tim Urban has a classic article in his Wait, but why blog about Your Life in Weeks. In it, he introduces this Life Calendar, a grid of 4,680 squares organized in 90 lines of 52. Each one represents one week of a 90-year life.
If you click on the image below, you can buy the poster. I don’t own it because I don’t believe in conserving physical objects for a lifetime. If you purchase it, I don’t get anything from it; I link it because it’s an interesting concept.
It takes many years to change the world (or a life)
Ambitious plans span much more extended periods.
Studying, starting a family, building a career, learning a skill, working for a company, or creating your own, are all multi-year journeys.
At Automattic, we use three-year plans to guide our strategy, and they serve a much longer-term vision. I look at my career as three distinct 15-year periods, the first of which I completed, with the third one starting in about three years.
I have a 10-year personal plan. Two daughters, too, each one representing roughly a 25-year commitment.
This is not to say that everything has to be planned. And long-term plans can and do change. But one year is way too short to be meaningful in the context of a life.
All this doesn’t mean that months, quarters, and years are useless.
They are reasonable timeframes to look back, review, and report.
🗺️ Three places to visit today
I believe 2023 will be a pivotal year because of technology, not just in the technology world. Here are three ideas to support this view.
I mentioned ChatGPT a few times already. I shared both the fascinating aspects and the flaws and tradeoffs. I think we are underestimating AI. It’s easier to focus on the problems and criticize what is still a prototype than face the reality that while the current tools are nowhere near general AI, neither are the requirements for many jobs. I recommend reading ChatGPT and How AI Disrupts Industries.
The current generation of Large Language Models requires enormous computers, and the next ones will need even more power. For the time being, building and running such machines remains limited to a few companies, which is why everyone uses the same models. Reading how This AI Supercomputer Has 13.5 Million Cores—and Was Built in Just Three Days opened my eyes to the fact that until now, AI research has had to do with whatever CPUs or GPUs were built for other reasons, as it was just research. With AI becoming a real productivity tool, the cost to train and run models will crate. Apple already announced Stable Diffusion with Core ML on Apple Silicon, meaning that soon we’ll be able to run (but probably not train) such models on our laptops.
Speaking of Apple, the company that put a powerful computer that we still call a phone in everyone’s hands and pockets is expected to launch one or two Augmented Reality or mixed Augmented/Virtual Reality headsets/glasses this year. I think this is the missing piece to AI becoming ubiquitous in the workplace, as many jobs where AI can help people are not done in front of a computer or don’t offer the possibility of having one in hand at all times.
🚀 It is rocket science!
Something is fascinating in the technology from the ’60s, from radios to TVs, from cars to trains, and even to rockets: it was advanced enough to represent incredible progress, to the point that some achievements from that era remain astonishing today, but simple enough to be understandable when explained.
Although I am passionate about today’s most modern devices, I miss being able to understand machines by just disassembling and looking at them.
If you have a couple of hours to spare, the next time you find yourself contemplating the bottomless depth and inanity of Netflix’s back catalog, I suggest you watch this video instead.
If you want a quick overview before watching, check this thread.