2023 is finally over which means it’s time for my annual stocktake of things I enjoyed throughout the year. If you’re new here, the standard disclaimer applies: I started doing this in 2020 after being inspired by a Robert Krulwich Radiolab episode. It’s not a ‘best of’ list, it’s simply a way of keeping track of what I chose to spend the most time with, what stuck with me or influenced my thinking, and what I recommend to others. As an added bonus if you read far enough down you'll also get a few thoughts on the year and what’s ahead.
If you end up liking this and want more, you can find previous years here: 2020, 2021, 2022. Thanks for reading.
TV, Film, Video
The Last of Us (HBO) - I had very high hopes for this video game adaptation going into the year; fortunately, it exceeded all of them. Episode 3 was probably the best hour of television I saw all year.
The Bear, Season 2 (FX) - It took a beat for this season to warm up but, good god, once it hit its stride. Episodes 6, 7, and the final were all particularly outstanding, with 7 (Forks) being my favorite.
Succession, Season 4 (HBO) - After loving the 2 pre-pandemic seasons I was pretty underwhelmed by Season 3: Greg-maxing fan service (“strong enough… for a man”) serving as the major crime. Thankfully this season course-corrected significantly. I did find The Tom Turn a bit jarring - while the decision to pull the plug early deserves applause, his abrupt character evolution felt like a tell they didn’t start out knowing where they wanted to end up. All in all, though, a satisfying ending and in a show full of unparalleled dialog, I think “Nothing happens in New York that doesn’t happen everywhere else” was my favorite line across all four seasons.
Slow Horses S3 (Apple TV) - Continued to be fun in its third season without getting too ridiculous. Short episode seasons leave you wanting more.
Only Murders in the Building, Season 3 (Hulu) - As a former Upper West Side dweller this show hits right from the main theme. I thought the shtick might wear out three seasons in, but just the right amount of new characters, new contexts, and self-awareness kept things from getting old. Coochie-coochie-coochie-coo.
Bluey (ABC) - Yes it’s an Australian kids show about a family of cattle dogs helpfully imprinting antipodean culture on my American children. It’s also a nuanced source of parenting inspiration that demonstrates how anything (7-minute episodes of a publicly funded TV children’s show in this case) can be a platform for excellence in craft and creativity if you try hard enough.
Jury Duty (Amazon Prime) - Did not know what to expect going in besides the obvious (it’s a reality TV show about a jury) and laughed the whole way through.
Yellow Jackets (Showtime) - Binged both seasons after hearing people rave for years. Strong Lost vibes with everything that entails. Season 2 was weaker than 1, but 1 was good enough to still have me looking forward to the third.
World Cup Sports (ESPN and Peacock) - This year was crazy for World Cup competitions in both cricket and rugby. And while watching the All Blacks lose to South Africa in the Rugby World Cup final was not the ideal way to celebrate my birthday, watching them beat Ireland in the quarter-final was, frankly, better than winning the title.
The Killer - Way more style than substance - but great on style and what’s a movie but a short story set to music these days anyway?
The Covenant - I watched this Gyllenhaal / Guy Ritchie collaboration on a plane expecting a dumb action film. Got all the dumb action I wanted but plenty more as well.
Barbie - Would I rewatch this? Probably not. Am I better for having seen it? I think so.
Useful Charts - A YouTube channel from a guy named Matt Baker who uses charts and data visualization to explore history and religion (and sometimes more contemporary/pop culture topics as well).
David Does Crosswords - My brother-in-law started this channel this year as an ASMR-y take on daily NYT crossword solves and it has become a regular presence in our living room. Just as highly recommended for non-relatives as well.
The Deal Guy - From his trademark open to his weird parking lot taste tests and undecipherable categorization methodology (if anyone knows the rationale behind how he groups his recommendations please let me know), this guy’s monthly ‘What to buy at Costco’ episodes are keenly anticipated in my household.
A bunch of things that were good enough that I watched all or most of them or thought were notable: Black Mirror S6 (good, actually) Fleishman is in Trouble, Foundation S2, The Crown S6, The Other Two S3 (a bit sad), A Murder at the End of the World (probably better as a YA novel), Extrapolations (overcooked), Eat the Menu (too much, but watching the Cheesecake Factory episodes as a three-episode run and nothing else is chef’s kiss).
Didn’t get to watch but excited to check out: Mrs Davis, Perry Mason S2, Drops of God, The Curse, Dumb Money, and The Holdovers.
Music
This year I found myself going deeper into the catalogs of artists I’d fallen for in recent years but never explored properly. Chief here was John Craigie, whose Opening for Steinbeck album was on a loop all year long. Lucero also got a fair amount of play. Honorable mentions also to Dessa, Corb Lund, The Beths, Little Simz, Boygenius, and Alex Lahey.
I made a list of the tracks I liked the most this year, as well as quarterly editions for Q1, Q2, Q3, and Q4.
Books
The Culture Series, Iain M Banks: I went deep on this series this year and feel like I could spend my whole life in here. Banks finished 10 separate novels set in ‘The Culture’ - a civilization governed by super-intelligent ship Minds - before his untimely death in 2013. Apparently quite popular with he who shall not be named, I wish this series (and other imagined futures from sci-fi authors generally) was way more present in today’s discussions about AI as a vision to work backward from. Each book is independent of the others and the series can be read in any order, although fans have thoughts on the best way to do it.
Ghost on the Throne: The Death of Alexander the Great and the Bloody Fight for his Empire, James S Romm: I continued to feed my interest in Alexander with this exploration into what happened immediately following his death. Better than anything in Game of Thrones and would make amazing source material for a prestige TV drama if anyone’s out there taking suggestions.
The Coming Wave: Technology, Power, and the 21st Century’s Greatest Dilemma, Mustafa Suleyman: More to come on AI below of course, but in a year where things accelerated faster than ever, this book is as good an overview of the key debates, ideas, risks, and opportunities as any. Does include some gratuitous axe-grinding and whining (complaining about technology pessimists, point-scoring on superintelligence concerns, etc) but not enough to undermine the overall value of the book. Touches a bit on biotech as well, which has not received as much attention as it should amidst all the AI hype this year.
The Creative Act: A Way of Being, Rick Rubin: There’s probably an argument to be made that Rubin’s interviews (this one with Tyler Cowen as an example as well as his own podcast) are as good if not better than the book itself, but at a time where it feels like human creativity is on the verge of being made redundant, it’s hard not to feel inspired by Rubin’s approach to work and life.
A World Without Work: Technology, Automation and How We Should Respond, Daniel Susskind: I picked this up when it was published in 2020 and decided to reread it in the context of this year’s advances in AI. Even more relevant and not any less disconcerting today than when it first came out.
Podcasts
Podcasts have become the principal way I stay engaged in what’s happening in the fields I care about. Below are the shows I listened to the most this year, as well as some of the episodes I found most interesting.
Shows:
AI and the future: 80,000 Hours, Dwarkesh Patel, The Future of Life Podcast, Build the Future, Lex Fridman.
Politics and the news cycle: Politics War Room with James Carville and Al Hunt, The Josh Marshall Podcast, The Bulwark Podcast (only the Friday episodes with Tim Miller, really)
Business and media: Pivot, The Rebooting Show, the China in Africa Podcast
History and more: The World Next Week, The Ancient World, Ancient Warfare, the Bart Ehrman Blog Podcast, Sound Opinions, BYC Podcast, Top of the Order, Unexplainable
Favorite Episodes:
Unexplainable: Can We Talk to Animals?
Search Engine: Does Anyone Actually Like Their Job?
80,000 Hours: Ajeya Cotra on Accidentally Teaching AI to Deceive us, Jan Leike on Super Intelligence Alignment, Tom Davidson on impact timelines, Ezra Klein on Existential Risk and What DC Could Do About It, Michael Webb on AI’s economic impact, Holden Karnofsky on how dumb AI could take over the world, Robert Long on Artificial Sentience
Dwarkesh: interviews with Paul Christiano, Shane Legg, Carl Shulman, Ilya Slutskever, and Nat Friedman,
Lex Fridman: interviews with Sam Altman, Jeff Bezos, and Neil Gershenfeld
Writing
Everything Richard Ngo writes on his site and on Twitter - like this thread on localizing status games and AI superorganisms
Sam Altman’s Tweets and What I Wish Someone Had Told Me
Ajeya Cotra at Planned Obsolescence
Paul Graham: How to Do Great Work
Elon University / Pew Research: The Future of Human Agency
Ethan Mollick: Setting Time on Fire
Joon Sung Park, Joseph O’Brien, Carrie Cai, Meredith Ringel Morris, Percy Liang, Michael Bernstein: Generative Agents: Interactive Simulacra of Human Behavior
Marc Andreessen: The Techno Optimist Manifesto
Matt Yglesias: All News is Bad News, The Techno-Optimist’s Fallacy
Aaron Shapiro: Is Your Product Better, Cheaper, or Easier?, Navigating the Commoditization Conundrum, Hypertargeting at Scale, The Death of the Big Idea
Everything Lulu Cheng Meservey writes at Flack.
Everything Simon Pulman writes on LinkedIn.
Charles Duhigg: The Inside Story of Microsoft’s Partnership with OpenAI
The NYT on Hot Ones and How to Solve a 25 Story Rubik’s Cube
Clare Ansberry: The Age When You Stop Feeling Young (this one hit hard)
Eugene Wei: How to Blow Up a Timeline
Seth Rosenberg: Product-Led AI
Nathan Baschez: A New Kind of Startup is Coming
Emily Wengert: The AI Agents Are Coming
Yoshua Bengio: How Rogue AIs May Arise
Joe Carlsmith: Predictable Updating About AI Risk
Tim O’Reilly: The Alignment Problem is Not New: Lessons for AI Governance from Corporate Governance
Maggie Appleton: The Expanding Dark Forest and Generative AI
Partnership on AI’s Synthetic Media Framework
Robert Long et al: Consciousness in AI
Abhilash Mishra: AI Alignment and Social Choice
OpenAI: Weak to Strong Generalization
Belinda Z. Li, Alex Tamkin, Noah Goodman, Jacob Andreas: Eliciting Human Preferences with Language Models
Threads
Websites, apps, and tools
ChatGPT and OpenAI’s GPT marketplace
The NYT’s new Connections game
Pyn’s open-sourced comms planner
Products
Uniqlo’s stretchy “Smart Ankle” pants - gotta get these in all the colors.
Taco Bell’s spicy potato soft tacos - the cheapest thing on the menu is also…the best?
Athletic Greens - I don’t know if this is really good or good for me, but I do like the discipline in getting my win in each morning.
Recipes
Alison Roman’s Steak Like Tartare
Molly Baz’s Spicy Feta dip
Sam Sifton’s No Recipe Recipes (also available as a book)
Year in Review, Thoughts on the Future
Toward the end of last year, I took on significantly more responsibility at work. It’s given me a unique and privileged vantage point from which to try to make sense of how the world is changing. It’s also occupied more of my mental, physical, and emotional capacity this year than any other. As a result, I’ve had less time to think expansively about the future than in previous years. But perhaps that doesn’t matter that much as this was the year the future I’ve been thinking about for a while started to turn into the present.
On the communications front, the most interesting thing to happen this year was downstream of the decision by Sam Altman and OpenAI to release ChatGPT 3 in late 2022 to the public to show people what AI was now capable of. This spurred a level of regulatory attention and civic engagement in 2023 that moved debate about everything that matters forward significantly and which would have otherwise come far too late to be useful. In an industry lacking in its own meaningful technical innovation, this kind of bold, action-based approach to comms has been the only development I’ve been excited to talk about this year. (Intelligent comms agents and the automated information revolution that is around the corner will change that soon, however.)
I suppose it’s also important to underscore the continued fragmentation between the perception of influence (influence with aging elites) and actual influence (influence with scaled audiences). One of the best things about working in Media is the cold hard reality of audience that we deal in every day. For me, this has reinforced the continental drift-sized divergence between what many people think matters and what actually matters at scale. Yes feeds and realities are more individuated than ever, but massive cultural moments and shared experiences now happen every day with highly informed people being completely unaware of their existence. I think in time this will create inverted Mandela Effects - memetic reference chains that will be obvious to millions but existed in parallel timelines to millions more and never existed at all for many people who are highly engaged in traditional information and culture ecosystems. Perhaps, after a brief period in which we thought we all plugged into the same reality - through broadcast TV, broadsheet news, and one or two primary newsfeeds - this is just a reversion to our historical norm. The more things change the more they stay the same, etc, etc.
The volume of information we are all asked to process now though is overwhelming - and that goes for both low-quality and high-quality information. The amount of opinion and analysis on almost any topic is more than one person could hope to digest - and that’s before LLMs accelerate the production of information even further. This unmanageable deluge has reinforced Aaron Shapiro’s commodification framework in how I now think about most things comms and marketing-related, from information production and consumption to broader marketing and business strategy.
But as mentioned, what really mattered this year were the significant advances in AI and the reaction and public engagement with those advances. This has been both surreal and exciting to watch, from the leap forward in capabilities themselves to the new applications in biology and material science, as well as the attempts at formal and self-regulation.
What else happened this year? Much of AI ethics was replaced by AI safety considerations, and then of course AI safety became its own battlefront in the EA vs E/acc wars. I’ve given a fair amount of thought to these issues in my spare time over the last six to seven years and on balance I’m encouraged by the number of people at the labs doing the most to advance AI who are giving way more than lip service to the issues. Will it be enough? I don’t know - but the state of the discussion is light-years ahead of where it used to be. And while I’m personally sympathetic to concerns about having one shot to align AI or risking lights out for humanity, I think we will have to get comfortable with a) humans ultimately not being in control and b) AI systems using a system of governance similar to our own to manage debate and complexity rather than deliver the perfect answer every time.
Ultimately I think it’s useful to conceive of humanity as a single self-managing global organism. Our world is full of pain, uncertainty, and injustice as people crash into each other’s interests and competing value systems. We mediate that chaos every day through debates, fights, the application of established systems and frameworks, and then the evolution of those systems and frameworks. I think the same will end up being true for how intelligent systems learn and evolve and negotiate with themselves - getting it right after getting it wrong, the same way we have for millennia.
Of course, we have a closing window for figuring out how to ensure people remain in an advantaged position given these developments and despite the engagement of politicians in the regulation of AI this year, the level of leadership and vision on the issue remain next to non-existent relative to what’s required. Absolutely everything can and should be re-evaluated through the lens of what AI can and should make possible to enhance the human experience. Where political vision is lacking, however, imagination in fiction is rich and I hope in the coming months we’ll see more attempts to index and extrapolate positive and constructive visions for the future from the decades of science fiction we have to guide us.
What comes next? I think in the immediate future, the best metaphor is probably the internet itself. We have been operating alongside a messy proto-superintelligence ever since we decided to daisy chain our little human brains together to form the global consciousness that is today’s social web. What happens next with AI will of course happen faster and be far more consequential, but I suspect the dynamics inherent in the internet era will likely be similar, at least at the outset of the AI era: things will get even faster (sped by automation and the elimination of manual human computer interaction and decision making) and be even more disruptive to economic and social institutions and norms. The same race that defined initial digital adaptation - large companies attempting to digitize their businesses before they could be properly disrupted by digitally endemic challengers - will now play out between fully digitized companies and new AI-endemic challengers. New services may well make existing businesses obsolete overnight but I think the biggest change will be the diminishment of humans in the economy overall: Business-2-Business commerce will become Bot-2-Bot as intelligent systems buy and sell from other intelligent systems and grow without needing people to run them. Business-2-Consumer commerce will become Bot-2-Bot-2-Consumer as people rely on intelligent systems to buy from intelligent systems on their behalf. Agents and dashboards will become foundational to how people manage their interaction with the intelligent world.
And what does all this mean for people? I do think we have some hard years ahead of us, but the opportunities are also enormous. For me, three tweets from three different people at OpenAI set it all out:
With the right leadership and vision, I’d like to think we can raise people’s expectations, exponentially improve quality of life for everyone, and - I hope - eventually remove labor from the human experience. Along the way we may have to come to terms with the realization that while human intelligence got to where we are now, being smart may not ultimately be what makes us special. This is a very difficult idea for people like myself who have built identities around being perceived as smart to get their brains around however, and I suspect we’re going to learn it slowly and painfully.
Before I end this post, I do want to make note of a couple of personal highlights for the year. I traveled more this year, including trips to Amsterdam, London, Cannes, and twice now to New Zealand. In February I somewhat miraculously made the last flight into Auckland before Cyclone Gabrielle demolished the east coast of the country to attend my grandmother’s funeral. And in December, my wife and I were able to bring our 2-year-old out to meet my family for the first time and I was also able to spend my first Christmas at home in 20 years. We also took my now 14-year-old twins to New Jersey and New York over the Thanksgiving break on our first full family trip post-pandemic. It was a really special time and re-confirmed for me that as wonderful as kids are when they’re children, your relationship with them as teens and adults can become even more special and meaningful.
Also, as I’m writing this note from a beach in New Zealand, I can’t finish it without commenting on things here. In another example of inspirational leadership at the start of the year, Jacinda Ardern stepped down as Prime Minister, ostensibly because she’d exhausted the energy necessary to run for re-election. I suspect also to give her party the best possible shot at winning re-election given the state of her favorability at the time. While it wasn’t enough, ultimately, to get the win, it was good to see her go out on her own terms and, hopefully, establish a precedent that will allow others to act similarly when their own time comes. New Zealand finds itself grappling with the kind of trying times that you’d expect a country of five million people to find themselves in at the moment amid a changing and challenged global economy and house wealth-driven inflation. My hope for home is that we’ll apply the same kind of bravery and boldness that has shaped our cultural history to shaping our future as well. Companies like RocketLab are demonstrating what it’s possible for Kiwis to achieve when we apply our irreverence for the possible and expected to solve our own problems and create our own opportunities. Technology companies routinely use New Zealand as a test market for their own new products and experiments, and I hope we’ll apply those same lessons to ourselves: this should be the little country that does big things first, and then scales them to the rest of the world.
Well, that’s it for 2023 version of this note. If this post is anything to go by, I expect the amount of change we’ll see in 2024 will be even more significant and that’s before we get to the 60+ different countries going to the polls over the next 12 months. I’ve not made any significant resolutions this year beyond the standard health and financial promises I make to myself every year, but given everything I’ve just written, this seems like a good one: less time spent trying to know everything myself and more time spent using the incredible new technologies now available to us to make more things for myself and others, while figuring out what it means to be human in an intelligent world.
That said, however, I wouldn’t mind learning why Orcas are (finally) attacking people. Maybe now that we can talk to whales they can tell us - although I’m not sure we’re going to love the answer when it comes.