Oct 26 | The Left Hand of Darkness by Ursula K. Le Guin (4/5)

A sci-fi story about a spacefaring envoy on a planet where the people are neither male nor female, but ‘ambisexual’: they fluctuate between genders on a monthly cadence. This is the most famous aspect of the book, and informs the story from fun details like the king’s pregnancy, through to deeper questions about the tensions that arise from our gender identity and expression. For a book written in 1969, it remains remarkably relevant today.

Like many sci-fi classics, The Left Hand of Darkness benefits from excellent world-building, but Le Guin has also written a great story on top of her world. The slow development of the two main characters drives the plot. As events unfold, they move from mutual distrust to profound camaraderie in the face of political scheming, faceless government, and a harsh, barren environment.

It’s an absorbing book; I read The Left Hand of Darkness front-to-back in a single sitting. Would recommend.

Sep 26 | Other Minds: The Octopus and the Evolution of Intelligent Life by Peter Godfrey-Smith (4/5)

A philosopher discusses consciousness by diving into the world of cephalopod intelligence, including octopuses and cuttlefish. Why cephalopods? Peter Godfrey-Smith likes them because they’re our most distant relative which still exhibits signs of intelligence — humans and cephalopods evolved intelligent brains on parallel paths. That comparison raises questions about our conception of intelligence.

For example, in primates there’s a correlation between brain size and how large the animal’s social network is (we have the biggest at up to 150, chimps 50, gorillas 20). But octopuses are largely solitary. So why get clever?

There are plenty of interesting diversions on perception and consciousness. Like, many people might know that cephalopods can change colour to camouflage themselves, or communicate to other animals. But how’s this? They’re colour blind too. We don’t even know for sure how they perceive colour to hide themselves — maybe the shape of their eyes induces chromatic aberration, or the skin itself can detect colour.

On consciousness, Other Minds discusses the idea of distributed intelligence. The analogy Godfrey-Smith uses is that of a rowing boat, with the crew swinging in sync with one another. Here’s an example: most animals’ brains have two halves. But humans are quite exceptional in how interconnected they are. If you teach a pigeon to solve a puzzle with only the left half of its brain, the right half doesn’t learn how to do it (you teach half a brain by covering up the eyes, apparently). So what is the perception of that pigeon like, if there isn’t a reconciled view of its surroundings?

Similarly, cephalopod neurons communicate more slowly than ours, so a lot of decision-making is distributed throughout the body. Their central brain doesn’t tell the arm to move; the arm moves, and lets the brain know it’s happening (via an ‘efference copy’ of the command to move). What if some of what we tell ourselves we do, is actually this kind of process in action? How much of our conception of a central conscious self is just a fiction constructed by different parts of our brain working in concert? Much to ponder…

Lastly, there’s one line that really stood out to me. There’s a drive in robotics to work on VLAs — basically LLMs like ChatGPT which have been tweaked to interact with the world. Here’s Godfrey-Smith on the role of language in human cognition: Language is an important tool for thought, and inner speech is not mere mental-acoustic froth. But it is not essential to the organisation of ideas, and language is not the medium of complex thought. That matches my own intuition — language is an important part of our intelligence, but it’s not the whole thing. And there’s a lot of physical intelligence exhibited by animals with hardly any language at all. That’s not to say VLAs are a dead end, but I do think there’s something missing.

In sum: it’s a good book, great food for thought, lovely images, would recommend.

Sep 26 | Surfing Uncertainty by Andy Clark (2/5)

The precursor to Clark’s book The Experience Machine. More technical, more detailed, more difficult.

Aug 26 | Consider the Lobster by David Foster Wallace (4/5)

A collection of essays covering a broad range from the Adult Video Awards, the crushingly mundane biographer of a tennis pro, and a deep dive into the rage of US talk radio. It was my first time reading anything by David Foster Wallace, but I can see why he garnered a reputation for phenomenal writing. The book is at once witty and a little melancholy. Strong recommend.

And while you’re here, listen to David Foster Wallace’s incredibly commencement address This Is Water.

Aug 26 | Infinity Machine by Sebastian Mallaby (3/5)

A month ago, Anthropic’s CEO, Dario Amodei, tweeted ‘At this point, saying that AI will cure cancer is more a cliche than it is inspiring, and most people think it is deceptive. The thing that will work is actually curing cancer.’

Demis Hassabis — the subject of Infinity Machine — is the only CEO of an AI lab that can credibly claim he’s actually made progress in that direction. His AI lab, DeepMind, released the protein folding model AlphaFold, as well as making contributions in other domains like weather prediction and nuclear fusion. As someone in research, I think it’d be hard to do much better than Hassabis.

The book runs through Hassabis’s life, charting his childhood as a chess prodigy, his discovery of AI as student, and later wheelings-and-dealings in the halls of power at Google. Over time, though, the focus sags. Mallaby has researched the book over the course of many interviews with central figures in modern AI. There’s a lot of information, but he crams in everything, even if it distracts: You get Elon Musk’s tantrums at OpenAI, Sam Altman’s ouster, Mustafa Suleyman’s departure from DeepMind, John Jumper’s time at DE Shaw Research, and more…

The other main issue is that the book barely interrogates its subjects. Hassabis says he’s only interested in scientific truth, not power or money. That may be true, but he does seem to have accrued a lot of power and money along the way — only lightly probed. He also has a clear competitive streak that may open up some blind spots, which Mallaby could have dug into more. Now that he’s stepped back from DeepMind, the proof will be in what he does next. Hopefully something a bit more interesting than LLMs: ‘Right now everyone’s just caught in the frenzy of “can we make the foundation model 10% better?” The other things we should be think about are receding into the distance.’

And anyway, I told a fib. The greatest sin committed in Infinity Machine isn’t that it lacks focus, or is a bit of a hagiography. It’s that the writing peaks at functionally dull and plummets to outright irritating. My Mum walked past while I was listening to it in the kitchen and asked what that ‘boring load of shite’ was. And worse, Mallaby has invented this stupid, grandiose term for AI which he insists on using constantly: the ‘infinity machine.’ The book is all ‘infinite this’, ‘infinite that’, ‘an infinity of possibilities’. Does it convey anything useful? Maybe only the infinite patience Mallaby expects of his reader (hahahaaaaaa zing!).

Anyway, in conclusion: the book is fine. I doubt many people would make it through this book without a prior interest in the subject. I got through it because I’m a fan of Hassabis. But in general, I’d say it’s just fine. Get it if you’re interested in Hassabis, AI, or Silicon Valley entrepreneurship. Skip otherwise.

Sidenote: this was my first time trying an audiobook (since you get one free one a month on Spotify). I’m split on whether I liked it. I’m sure I’ll retain less of this book than I normally would if I’d read it off the page — it’s so easy to drift away or miss a detail when you’re also doing things in the real world. But it was nice as a sort of extended podcast, for a topic I didn’t care about too deeply.

Aug 26 | The Art of Explanation by Ros Atkins (2/5)

I was disappointed with The Art of Explanation. I had hoped for a discussion of how explanation works and how we actually get the ‘Aha!’ moment — something to fire a bit of imagination. Instead, the book is practical, and a bit mundane. It’s dedicated to the recipe that Atkins uses to write articles or scripts. That means: the features of a good explanation in numbered list form; a multistep formula to be applied when planning an explainer; grey marks on the sides of the pages so you can open it back up to step 4 like a reference manual.

I found the book at once too-detailed and hard-to-follow — ideas like “use simple language” occupy multiple pages, but it felt like it was explaining the bleeding obvious without many useful examples of what he’d consider simple. In the end I skimmed the book at a growing speed over the course of a two hour train ride. I don’t feel like I need to revisit it to fill in the gaps.

That’s not to say the book is worthless, or that the ideas don’t make sense. I’ve tried some of Atkins’ tips; like organising information into ‘threads’ and testing different orders before getting into the writing. They’re handy. Even so, I can’t shake the feeling that, in the same way that there are meetings which could’ve been an email, this book could have been a blog post.

Aug 26 | Everything I Know About Love by Dolly Alderton

A cozy autobiographical read touching on self-worth, relationships (both platonic and romantic), and life as a young millennial in London. The book was widely popular, but I feel like I’m in a particularly tight circle of readers. Alongside her misadventures in love, Alderton leads a certain kind of London posho path through life (private school -> uni -> journo) that I recognised — I come from the same set. Honestly, she might as well have been my older sister Cat (who even recommended this book to me in the past). The familiarity meant that Alderton’s jokes and observations all landed well for me. I thought this book was pretty funny.

Strangely, though, the connections to my own life extend beyond that. Everything I Know About Love is dedicated to, and a chapter is given over to describing, a friend (and late-teenage crush) of mine — Florence Kleiner. Flo was funny and charming and energetic. Her chapter jogged memories I’d not thought about in years. As seventeen-year-olds we saw Matthew Bourne’s Swan Lake with some friends. During the final climax, the swans brutally savage and attack the lead as he tries to protect his beloved Prince. I looked over to Flo, and saw tears welling in her eyes, spilling over, and quietly trickling down her cheek.

Flo passed away from leukaemia when she was 19. Whatever the writing or the narrative or the tone, this book gave me a chance to remember her. It’s a rare thing, and I’m grateful for it.

Jul 26 | The Lost Homestead by Marina Wheeler (3/5)

The Lost Homestead is tough to pin down — it’s all-at-once a memoir, a biography, and a brief history of India. The book covers the run up to Partition, when the country split into Pakistan and India, through to the era of the current Modi government. Wheeler digs into the history and shares stories of her mother, Dip. Dip is a likable, self-possessed main character; when she leaves India with her English husband, Dip’s mother worries that, without servants, she’ll spend her life scrubbing pans. In the present, Wheeler looks up from her laptop to see her mother’s hands in the sink.

Alongside the memoir, Wheeler offers commentary on Indian history and politics. I don’t know enough to judge, but these sections feel shallow. The biographical detail leans on her wealthy family’s history, who were shielded from much of the turmoil she describes. For example, the great loss of Partition involves her grandfather abandoning his estate in the Punjab, to move his family into a, err, glittering mansion in Delhi. When Wheeler wonders whether the British Empire is unfairly vilified, she marvels at the railways, but skims over the Bengal Famine.

I haven’t read much about modern Indian history. To that end, the book was an easy-to-read, broad strokes introduction with a charming personal touch. But it fell short as a history of India. I’d read it if you’re in the market for an easy-to-read story about a daughter and her mother with light detail on Indian history.

Jun 26 | What Is Intelligence? by Blaise Agüera y Arcas (5/5)

A fascinating read: thoughtful, dense (very dense) with technical detail, and quite provocative. What Is Intelligence? is fairly broad, covering topics from evolving code, to the decentralised nature of consciousness, and whether sequence-learning is sufficient for generally intelligent AI. I’ve never found myself wanting to reread a 600-page tome before, but there’s so much information in the book that I suspect I need another go-around just to begin absorbing it.

The book is contrarian. Early on, Agüera tells the reader that he believes LLMs like ChatGPT are intelligent. I balked at that — I reckon most people would. But through the book he works through his thoughts in such a considered, self-consistent manner that it really does bear fruit. I’m still sceptical, but far more sympathetic to Agüera’s ideas than I would’ve expected.

Beyond the question of AI sentience, the book is packed with interesting information. One memorable discussion touches on ‘sphexishness’ — behaviour that appears intelligent until you know how it works (named for a kind of wasp). Another part I enjoyed gets into the decentralised nature of cognition in octopuses, and how it relates to split-brain human patients. I’ve never read anything quite like it.

Ultimately, while I’m still sceptical of Agüera’s ideas about intelligence, his views are argued so clearly, and with such a wealth of supporting evidence that I’ve grown much more open to them. Either way, the book is a fascinating, challenging read. Well worth the time.

May 26 | Maintenance: Of Everything by Stewart Brand (4/5)

‘Taking responsibility for something can be a radical act.’

On the surface, Maintenance: Of Everything is about keeping sailing boats and motorbikes in good nick, but there’s more to it — how we maintain ourselves, our relationships and our lives. While I was reading the book, a magnet fell out of the case housing my e-reader (the cheapest one on Amazon). Previously I might have lived without, or bought a new one. Instead, I fixed it. Now I feel a small twinge of pride whenever I use the e-reader.

The book is beautifully designed and put together, and it includes interesting sidebars from authors that commented online as Brand openly drafted the book on Works in Progress. Strong recommend.

Apr 26 | Cat’s Cradle by Kurt Vonnegut (4/5)

Very short, very weird, very funny, a little melancholy, hard to describe. My only regret is that I gave it away, so I’ll have to buy it again for a reread.

Apr 26 | The Scientist in the Crib by Alison Gopnik, Andrew Meltzoff, Patricia Kuhl (3/5)

The Scientist in the Crib is a fairly short read by three Childhood Development researchers. The book describes how young infants learn to piece the world together. The identify three core aspects of childhood learning:

• Representation learning, driven by genetics (e.g. object-centric vision, distinguishing speech patterns)
• Learning via physical and linguistic experimentation
• Interaction with other people — both learning from them (that’s two-way — think about the exaggerated, easily distinguished mother-ese we naturally use with babies), and about them (e.g. mentalisation, Theory of Mind).

They hypothesise that we exhibit a curiosity drive that compels us to explain things we see, that we feel as powerfully as any other biological impulse (take your pick from the Four F’s — fight, flee, feed,… reproduce). One memorable example from the book — if you connect a baby’s foot via a silk ribbon to a mobile, they will learn to kick their foot to make it move. If you remove the ribbon, they’ll try to kick but find it no longer moves. Puzzled, they’ll resort to another trick — one that works on Mum — they’ll earnestly smile at the mobile, to see if that moves it. That’s while they’re still figuring out causality I guess. Either way, a good read, although I wish it had a little more scientific detail.

Apr 26 | On Writing: A Memoir of the Craft by Stephen King (4/5)

A great discussion of writing — think Strunk & White’s classic style guide, only with more depth and examples. It helps that King’s writing is so fun to read — it’s evocative, it’s conversational, it’s intelligent. He starts with a brief memoir that I would have happily read by itself. It’s followed by more specifics on writing, naturally with a focus on fiction (I reckon it carries over to nonfiction too though). The general thrust is to write lots, read lots, and be honest. The best part is the closer, in which he offers an unedited and edited version of the same passage — the second version comes alive. A fairly easy ‘recommend’ for anyone interested in the craft of writing.

Some memorable directions:

• No adverbs: ‘to write adverbs is human, to write he said or she said is divine’
• Active beats passive (‘Frank killed John’ beats ‘John was killed by Frank’)
• ‘You must do two things above all others: read a lot and write a lot. There’s no way around these two things.’
• ‘Reading takes time, and the glass teat takes too much of it.’ Lots of detail on setting up a workspace and limiting distraction so you can work hard.
• ‘What are you going to write about? Anything at all… as long as you tell the truth.’
• ‘The basic rule of vocabulary is always use the first word that comes to your mind, if it is appropriate.’
• ‘2nd draft = 1st draft - 10%’
• Let a piece sit for six weeks after the first draft, and do not touch it or look at it. It will feel like someone else’s piece when you return, and it is easier to kill someone else’s darlings than your own.

Mar 26 | Life Ascending: The Ten Great Inventions of Evolution by Nick Lane

Perhaps I was overly excitable after crashing through A Brief History of Intelligence, but Life Ascending isn’t for me. It’s mostly about the biochemistry of the origin of life, which although interesting, I didn’t find very inspiring — I ended up dropping the book ~150 pages in.

Mar 26 | A Brief History of Intelligence by Max S. Bennett (5/5)

The book breaks down the evolutionary innovations underpinning intelligence. It steps through five innovations, with parallels in AI: steering (RL), pattern recognition (ML), simulation of your own actions (world models), simulation of others’ state (Multi-Agent RL and Theory of Mind), and language (NLP). Quite a few research ideas struck me while reading the book.

As someone working on AI research, this book is a humbling read — the human brain achieves all the things we’d like to do simultaneously, efficiently, in a world with no resets. The more you learn about the brain and the body, the more AI and robotics looks like a poor imitation. Modern AI is improving rapidly, but we have no unified model for all five of Bennett’s innovations, and nor do we seem very close. Cannot recommend strongly enough, for anyone interested in the nature of intelligence. Thanks to Tim Röcktaschel for including a recommendation for this book in his UCL course on Open-Endedness.

Feb 26 | The Art of Doing Science and Engineering by Richard Hamming

In 1995, Turing Award winner Richard Hamming gave an incredible talk titled ‘You and Your Research’. In the talk, he laid out the patterns he had observed which lead to successful research careers — what enabled people to do good work. This book takes the lecture and stretches it out with biographical detail, covering his life’s work. Expect anecdotes about error correction, mainframes, and FORTRAN. Personally I dropped the book after getting about 2/3rds in — I don’t think it’s worth reading unless you’re interested in Hamming’s era of computing. The lecture, however, is essential viewing.

Jan 26 | The Siege by Ben MacIntyre (3/5)

A gift from Dad, not the sort of thing I would pick out for myself. But it’s a good romp about the Iranian hostage situation in London (not to be confused with a similar story at the US embassy in Tehran). The book gives some context on the Iranian Revolution and subsequent downfall of the Shah, which has unfortunately turned out to be quite relevant at the time of writing, but adds some great colour. I preferred the first 2/3rds of the book, which covers the situation within the embassy. As the tension ratchets up, it’s hard not to feel some sympathy for the hostage-takers: as their situation worsens, it becomes increasingly apparent that they’re in far over their heads. The final third details an SAS rescue attempt, characterised mostly by total chaos. A nice page-turner.

Dec 25 | Human Compatible by Stuart Russell (3/5)

A good read by one of the pioneers of Inverse RL, in which an agent tries to infer the reward function a human optimises for. In Russell’s framing, safe, aligned AI will need to truly understand complex human preferences. The risk of pure reward maximisation (as in standard RL) is opening yourself up to reward hacking. Inverse RL comes with its own challenges though: how do you handle human preferences that change constantly, whether due to uncertainty, or just shifts in what we value?

The book is well-written and clearly communicated, but for me it doesn’t fully address the question of capability. RL’s great successes have been agents which learn to exceed human capability. Doesn’t inverse RL cap you out? Think about AlphaGo, DeepMind’s Go-player pretrained on human data, vs. AlphaZero, which was trained with no human data. AlphaZero took longer to train, but its asymptotic performance exceeded AlphaGo’s. How much capability would we be willing to concede for the sake of a more human behaviour? Overall, a good read, would recommend.

Feb 25 | Why Greatness Cannot be Planned by Ken Stanley & Joel Lehman (5/5)

A discussion of invention and discovery, centred on a machine learning experiment named PicBreeder. The authors argue that interesting discoveries don’t happen because of straight shots from A to Z; rather, intermediate results that enable discoveries are unpredictable, forming stepping stones to new ideas. An example from the book: Suppose you’re some God three billion years ago, and you want to develop an intelligent species from the primordial soup. Should you start administering the SAT to single-celled gunk in the hopes of breeding big-brained bacteria? Probably not. To develop intelligent life, many apparently-unrelated steps had to happen first: meat-eating, warm-bloodedness, aerobic respiration, eyes, lungs, legs, and more.

Stanley and Lehman’s point, then, is that we should focus on novel or interesting discoveries which are within reach, rather than aiming for distant goals which we can’t even sketch a route to. In ML terms, they advocate evolutionary methods and novelty search over objective-driven optimisation. Beyond ML, they argue that we should embrace serendipity and permit ourselves to follow our interests. A fantastic book which altered my perspective on decision-making in both ML and real life.

To Write Up

• Oct 25 — The Experience Machine by Andy Clark (4/5)
• Jan 25 — Thinking, Fast and Slow by Daniel Kahneman (3/5)
• Feb 25 — Abundance by Ezra Klein (4/5)