Happy New Year! I hope you've had a great holiday period and new year celebrations, wherever you are.
At the end of 2022 I wrote some reflections on a year of being a “solopreneur” and moving to Germany. In 2021 I wrote about leaving government and starting my freelance business. I’ve found these annual exercises useful, in particular by comparing to whatever I wrote the previous year.
So here is this year's post. It will be a more introspective one across a few themes – finding stability, job-hunting and job-finding, and some thoughts on the state of the world in 23/24. And then I’ll end with some recommendations about habits and board games, and some cartoons to send you into 2024 with some amusement.
I (finally) settled in Berlin
Bureaucracy can be very emotional. That might sound like I spent too long in the Civil Service, but bear with me. One of the biggest moments for me this year was getting my German visa in October. For context I’d moved to Berlin in February 2022, i.e. more than 20 months earlier. For some of this I was able to travel normally (ish)* but from Feb-Oct this year I couldn’t leave Germany. At one point, due to bureaucratic miscommunication, I thought I was about to be deported (I wasn’t, though it took three days of emailing to establish that).
There are other clear low-points I can speak to – e.g. even once my visa was confirmed, I had to wait 6 weeks for it to be delivered (I missed two weddings in the interim). But I found the worst part to be the longer-term low-lying stuff, the uncertainty, the inability to plan, the feeling of lacking control over one’s life, the sense of trying to build a life in a place that isn’t letting you feel settled. There’ll be some examples in the other sections.
The broader, most important, point: even as someone with all my advantages (secure finance, English-speaking, PhD, stereotypically British name), I found the immigration process extremely stressful, and I really feel for those trying to do it without my advantages. Whatever one’s views on immigration – and I suspect debates we’ve already seen this year are nothing compared to what’s coming in future – I hope there can be some empathy for how bloody difficult the thing can be, and how the people going through it feel.
But back to positive news – I also found a long-term flat in Berlin. As Berliners will tell you, this is a monumental achievement (I’ve had 10 short-term places in 18 months). So, in a very nice comparison to this time last year, I know that I can travel back to Berlin, there will be my familiar flat waiting for me, and I won’t be worried about whether I can stay. Normality feels very nice.
* I had a jobseekers’ visa for 6 months, also some visa-free days that Brits get; so there was a lot of counting days involved, which was also stressful.
I found employment
Last year I wrote about trying to move from working ‘tactically’ to working ‘strategically’. For me, that meant trying to leave freelancing and:
return to work in an organisation with structure and full colleagues. The organisation will ideally work at some intersection of technology and society, and trying to do something socially good. It’ll have problems to solve, whether internal or external; I can use my different bits of research/comms/tech/interpersonal skills as needed in different situations.
The other major milestone for me this year was joining AlgorithmWatch – a human rights organisation “for a world where algorithms and AI do not weaken justice, democracy, and sustainability, but strengthen them” – to lead their Auditing Algorithms for Systemic Risks project, which involves a mix of research, network building, strategizing, and more. So I think I can tick the above 2023 goal off.
Getting this job was a fantastic moment for me – I’ve known the work of AlgorithmWatch for a while, and it’s been living up to my expectations. But it only came after a very large number of rejections over many months, some unsolicited and many which did not excite me at all (I did get very far with one exciting option, which then turned out to be unable to offer German visas. Back to the previous point!). So by the time I applied for AlgorithmWatch I had basically written off my chances of getting it, given this track record and how exciting the job sounded to me (and probably a lot of other people).
Something like this has happened every single time I have been on the job market; in particular when I was rejected from various public sector research jobs after my PhD, before getting a job in Number 10 Downing Street – which I’d applied to largely for interview practice in case that kind of job came up again a few years later. “Good things come to those who wait” is unhelpful advice for someone who is actually waiting, so I guess better advice might be: don’t calibrate your expectations and behaviour too much on failures, even if you learn from them.
I’ll also plug, as I did last year, Molly Graham’s excellent blogpost on this topic, on finding the “holes that are shaped like you”. I’d add that part of this is finding individuals who are like you – when you attend events, read a post you like, whatever – and then looking up where they work, rather than just thinking about organisations. Some of the places I applied to are still, to me, faceless entities that I will probably never think of again. But often the better application experiences started from, or ended up with, positive personal connections. Maybe the true job application is the friends we make along the way, etc. etc.
I tried to pull together some thoughts on the experience of returning to full employment, but it’s only been a few months and it’s all still settling in my head. The one which does jump out is, weirdly, similar to something I wrote about going freelance in 2021: “Announcing my plans brought instant kind (and free!) advice from contacts old & new. Now I’m part of various great networks where we make opportunities to work together – or just to share ideas.” This is also true of the AlgorithmWatch and the wider tech-policy community it is part of, who are proving an eminently driven, helpful, and enjoyable group to work with. I look forward to more in 2024.
Also to note – when I say “full employment”, I am with AlgorithmWatch 4 days a week so carrying on other freelance work. In particular I’m very happy to still be working with the excellent Institute for Strategic Dialogue, which is keeping me connected to the coal-face of social media research. I also had a great time building a course on digital research skills for Kiron, an excellent NGO which supports education for refugees. So if you have a project you think I could help on, please still feel free to get in touch. Public examples of previous things I’ve done can be found here: https://www.oliver-marsh.com/post/work-roundup
I wasn’t quite as strategic as I’d planned
As mentioned, a major theme of last year’s piece was moving from ‘tactical’ to ‘strategic’; in brief, moving from experimenting with lots of things to focusing attention on fewer things.
I definitely did follow the maxim of Michael Porter: “The Essence of Strategy is Choosing What Not to Do.” My decision to put my proto-business The Data Skills Consultancy on ice was the right one. My prediction that finding full time employment would help me focus attention seems to have been correct so far, to an extent which has surprised me. I’ve made headroads into a few things which have sat dormant for months – mostly Life Admin, but (probably more interesting to readers of this newsletter) also an ongoing side-project into Twitter/X Community Notes, which I hope to publish results of soon.
There is a weird psychological effect of being able to separate clearly “work time”, “side project and admin time”, and “social time” in a way I couldn’t do so easily when freelancing. When the time available for each is less flexible, I’ve found it easier to use each more wisely. (Also, gaining the background mental energy – and time! – previously expended on things like getting a flat and a visa has been a huge plus).
However. Linking “things I want to accomplish” with “external event happening” is a good approach – more on that in the “Fun Fact” below. Having decided I was trying to get a job, I basically treated the interim period as “do short term freelance projects, plus other personal projects that will be harder to do once fully employed”. But the approach is not so good when the event happens at some indeterminate point in the future; by the time I joined AlgorithmWatch, 2023 was already 2/3 over.
Periods of stasis can seem like good opportunities - like how many of us, myself included, saw initial Covid lockdowns as a chance to Do Some Interesting Things and then did not Do Those Things. In 2023 I started, but did not finish, some papers from my PhD; didn’t do as many of these newsletters as I would have liked; and I did not travel anywhere in Germany outside Berlin-Brandenburg, despite being stuck in Germany. Unsurprisingly, the most productive projects in both periods were those which had a deadlines (a weird essay I wrote for a competition during Covid, a presentation in German during 2023).*
The trick I missed was seeking out external deadlines. That was hard during the worldwide period of stasis in Covid; I definitely could have done that last year. Finding a job is hard and unpredictable, so finding some easier timepoints would have broken up the stasis a bit.
The positive side, as noted above, is that since being employed has meant I’ve seemingly become better at ticking stuff off. Ironically, since I’ve been employed I’ve visited Dresden and have plans to go to Karlsruhe. But still, it’s always been easy for me to say “I’m not doing x because I’m too busy”. I think the experimental results don’t back that up.
* More info: In Covid, I wrote a (strange but fun) piece for the Nine Dot Prize on Aging. In 2023, I prepared (with a lot of support) a lecture in German for the Humboldt university on using AI for clustering Twitter communities. Neither were “successes” if measured purely by outcome – I didn’t win the prize, and the Humboldt asked me to do the lecture in English. But both of them helped me focus on things – reading and writing more about politics, and improving my German – that otherwise I was bumbling around a bit aimlessly.
The World in 2023/2024
It’d be remiss of me, sadly, not to mention the state of the world as we enter 2024. This bit is depressing, so maybe skip if you’re not feeling great – we don’t all need to absorb all the available negativity.
Though the economic situation of 2024 seems (maybe) to be looking up – which could have additional knock-on effects, just as a bad economy helps negativity to flourish – there’s still lots to be concerned about. Next year sees elections covering, by some estimates, 3-4 billion people worldwide. The US is obviously the biggest, but also Taiwan, India, the European Parliament (with some deeply problematic parties seeing renewed popularity), and probably the UK (though I’m less worried about that one). I have previously argued against overconfident predictions about exactly how AI will affect elections, but I still think it can and will, probably negatively.
Also, October 7th and the ongoing reporting out of Gaza continue to be horrifying. The Ukraine situation looks more desperate than before. Things that once felt settled – downplaying of climate change, xenophobia, sexual and reproductive rights, democracy – feel worryingly up for grabs again. These are all bad in themselves. But a specific concern that occupies me, personally as well as professionally, is that all these provide opportunities for further division, and power for those who can best weaponise negativity.
A related concern is that those of us who want to stand up for those things – “progressives” seems to be the word of the moment – don’t have the language with which to convince people, having had years of treating these subjects as fairly settled. I originally moved from science communication to politics because I felt behaviour I was seeing in my PhD research on social media – people who saw themselves as “the good guys”, the intelligent and rational ones, publicly belittling and dismissing views of those unlike them – were also happening around Brexit, Trump, and western politics generally. I’m not that happens quite so much anymore. But I’m not sure we’ve got good messages which don’t rely on referring to “the far right” or “climate emergency” or “threats to democracy” and hoping those will scare people.
(But this may be unfair; if you disagree with that, if you think there are better messages, I’d love to hear them).
Polarised language happens on television and radio debates, in newspapers, in political speeches – and arguably that has more impact than social media (though of course social media is easier to blame for journalists and politicians). But social media, and other modern technology, is what I know best so I’ll comment on that. I find X/Twitter has degraded so much I don’t check it as much, which means when I do I notice the negativity of the experience so much more. I feel I’m starting to see that creep into LinkedIn, with proponents of differing views on tech, AI, and regulation having increasingly harsh words for one another. The much-shared idea of 4 “camps” of AI views is, in my view, becoming as unhelpful and damaging as dividing politics into ‘left’ and ‘right’. At a time where we most need some kind of defences, with powerful tools and real threats proliferating, the debate risks being distracted by negative stereotyping (e.g. of “regulation” as a bad thing in general) and hostility.
I'll end with something I can suggest more positively. Much as I love the field I work in, it does expose people – many get much more exposed than I do – to a lot of concerns. It also leaves lots of us feeling powerless, particularly when a small number of people – whether Elon Musk, Jim Jordan, or (in a different way) Thierry Breton – have the power to rapidly undermine years of progress. Plus, of course, I know people who are personally affected by events. And, simple as it is, I have found that a quick text checking in on people seems to make a difference. Sometimes it’s just gets a quick response, but other times it opens up bigger conversations. And, if the conversation leads to disagreements, the start-point means we can debate openly based on trust. And even just sending a text feels like doing something at least.
A theme running through many of my posts (especially this one about Twitter) has been that macro shouldn’t distract from the micro. It’s easy to get bogged down in talking about how social media, or AI, or whatever, might lead to the macro-scale downfall of democracy or civilisation. But it may be in the more targeted, micro-problems around particular individuals, groups, and events that we can more clearly discern the impacts of new technologies. But maybe micro, one-to-one, personal communication can be a positive force in 2024 too.
In the meantime, here's some researched lists of effective charities to donate to around Gaza (one, two). I have also found this charity against hate speech in Germany, but it's harder to find good research on effectiveness; any other recommendations very welcome.
___
Fun Fact About: Habits
New years bring new years’ resolutions. I’ve been reading Atomic Habits by James Clear, and I begrudgingly recommend it. Begrudgingly, because I find the style – a very formulaic “American CEO/entrepreneur gives life advice” – a bit grating, if still readable. But I’m already finding his suggestions surprisingly effective.
One tip which I mentioned earlier – connect a habit you want to develop, with a regular habit you already have. I have been trying, with mixed success, to have a daily routine of ‘dictate some German for 5-10 minutes, then use DeepL and ChatGPT to see the mistakes’. A habit I already have is eating breakfast every day. So I now do the German exercise straight after breakfast; nothing else after breakfast until I’ve done the German. And I’ve found that to be a real improvement.
Another tip is to arrange your environment to make things harder or easier. I have long been setting my phone to greyscale so it’s less fun to use. I’ve now also moved my ‘distraction’ apps (Twitter, Instagram) to harder-to-reach places in my phone, and apps I want to use more (Feedly for reading, Quizlet for German) to where my thumb naturally falls. I check the distraction apps less, but more progress needed on the other part.
Lots of the suggestions he makes seem very obvious. But it’s still been weirdly helpful to be told by an external voice to do these obvious things. And, as a bonus recommendation, I suggest this excellent book on social research Everything is Obvious (Once You Know the Answer) by Duncan Watts.
___
Recommendations
Podcasts: My go-to tech podcasts have appeared throughout previous newsletters, but here they are collected together: Decoder, Moderated Content, the Stratechery bundle (especially Dithering), Hard Fork, the VergeCast, Politico Tech, The Sunday Show from Tech Policy Press, and tech-focused episodes of Lawfare and Capitalisnt.
Book: Best book I read last year was The File by Timothy Garton-Ash. It’s a record of his experience reading the file that the Stasi compiled on him while he was researching and travelling in East Germany, compared against his own diaries and writings from the time, and then meeting the Stasi informers to ask them about their motives.
Games: Over Christmas a friend introduced me to the board game Twilight Struggle, a two-player game reenacting the Cold War. It’s complicated, but there’s a reason it’s won a lot of accolades from board gamers. If you like complicated board games, I’d also recommend Tzolkin: The Mayan Calendar. If you don’t like complicated games, I’d suggest Codewords as good teamwork fun.
Funnies: I’ve recently started following more cartoonists on Instagram, and they’ve quickly become one of my favourite parts of my feed. Two I’d particularly like to recommend are Nathan Pyle (Instagram here) and Sarah Anderson AKA Sarah’s Scribbles (Instagram here). Sarah Anderson is particularly worth mentioning in this newsletter as she’s part of one of the biggest lawsuits against use of artists’ work to train AI models without permission. I’ll end with two tech-themes pieces from these two. Happy New Year!
Thanks for reading. Please do share with others using this link, and let me know your thoughts via this short poll.