Antonym The LLMsip edition
Open the Lemsip sachet, Dave.
Dear Reader
Happy New Year. I’ll start with a grim personal anecdote, then wrangle us back to talking about AI, as usual.
I missed Christmas this year due to being a man-flu sufferer with tragic-actor/hypochondriacal tendencies. Histrionics aside, the depths of a winter-whatever-virus can be a dark, fearful place. Like many, when perplexed by illness I turn to the worst possible information source: Google.
Questions like “Is is normal to feel like a corpse after two days of a cold?”, “Dizziness and nausea cold symptoms” yield a crowdsourced diagnosis of certain doom.
Typically, after reading a couple of misspelled conspiracy comments on Reddit, some news source scare stories and some melee bullying on Mumsnet, I’ll be convinced something is cataclysmically wrong with me. Then I generally worry for a good half hour or so before the Lemsip kicks in, I feel slightly better and then start to binge-watching a Disney+ show about an Australian hitman with heart of ambivalent gold (see recommendations below).
Not this Christmas! Instead, I created LLMsip, an AI bot coach for the severely self-pitying sufferer from colds. (At the time I was testing negative for Covid and was convinced it was “just” a bad, albeit possibly terminal, cold.) The bot was actually pretty helpful in this context. It explained what was happening and sympathised and suggested the usual sensible things to do to get better.
You can try it here. (In no way affiliated with the Lemsip brand – I was just drinking a lot of at the time.)
Reminder: reading is good for you
After a few days of festive lurgy, I was able to concentrate enough to read books again.
A reading habit doesn’t look after itself. It takes focus and prioritising.
Reading, lest we forget is a healthy thing to do, as Margaret Heffernan reminded us in a piece for the FT:
Neuroscientists have been at pains to demonstrate that the pleasure a book provides isn’t indulgence; it’s good for you. Reading will keep you better informed about the world but it can also improve our tech-shattered ability to concentrate. Standing in the shoes of others fine tunes our social understanding, useful as we struggle to understand friends, neighbours, customers and co-workers. Different parts of our brain engage as we simulate scenes, characters and mental states. Our imagination — remember that? — is rekindled.
Reading has also been found to make us more helpful, to reduce bias, and even to increase longevity — something we will enjoy all the more if we have a good book in our hands.
Living longer may be correlation not causation, but the data is real. In The Art of Rest, Claudia Hammond notes that paradoxically, while reading is restful activity, it is intensely stimulating:
More than 3,000 people were asked how much time they spent reading books, magazines or newspapers during the previous week. Forty-one per cent read no books at all. Others were keen readers. The sample was followed for a decade and during that time just over a quarter of the people died, but the good news for the bookworms was that they lived an average of almost two years longer than the people who only read newspapers and magazines.
Big claims. One of the books I read was Going Infinite by Michael Lewis, an account of the brief rise and fall of Sam Bankman Fried, who started a cryptocurrency business that was careless to the point of criminality with its customers’ money (he’s now awaiting sentencing for this, somewhere in the range of 25 - 110 years). The weird logic of his extremist utilitarianism manifested as, among other things, not reading books, which Molly Roberts said in a piece for the Washington Post “tells you everything” about his lack of imagination and interest in the inner lives of others. Bankman Fried apparently said:
I don’t want to say no book is ever worth reading, but I actually do believe something pretty close to that. … If you wrote a book, you f---ed up, and it should have been a six-paragraph blog post.
Of course, he could have a point with many business books (see the reading rules below, especially the point about filleting books).
In his Lunch With The FT interview this week Italian author Giuliano Da Empoli says:
“What’s incredible about fiction is it takes you outside your head and inside someone else’s. It’s more and more difficult to get out of our own heads. Because everything reinforces what we already believe. That’s why the difference between fiction and non-fiction readers is immediately apparent when you talk to people. Fiction readers understand that theirs is not the only perspective. That each of the people sitting around us” — he points — “have their own. If you read fiction, you’ve been in some of those heads. If you don’t, even if you read non-fiction, you don’t have that experience.”
I buy that argument. And I look forward to reading Empoli’s fictionalised account of Putin’s reign, The Wizard of the Kremlin. As Hilary Mantel said about historical fiction, imagination jumps off from the facts and lets you explore the unrecordable inner lives of real people.
How to read more…
My current reading rules are:
If in doubt, set a timer and read for 15 minutes. Reading is a habit and sometimes it can get broken and needs a helping nudge. Sometimes reading can feel like a chore – this is a trick of the mind. Fifteen minutes means you read until you’re in the flow and then it’s hard to stop. But don’t force it.
Have books for different places – a bedroom book, an office book, a book for the bag, and one or two on your phone, maybe an audiobook. Sometimes a book is so gripping it has to be taken everywhere, other times a book is happiest in a certain place or time.
Non-fiction fillets. If you’re not sure if a non-fiction book will be worth reading the whole of - fillet it. Read the introduction and the conclusion, the start and end end of each chapter, skim the rest. Either you will get what you need this way or it will turn out to be so good you’ll read the whole thing anyway.
If you’re not enjoying a book: stop. You can throw it across the room in disgust, or just set it aside for when you might feel like returning to it.
Set a reading goal on Goodreads or whatever – but don’t worry about it. I missed my “goal” last year by about 13 books, but it doesn’t matter. Much more rewarding than hitting made-up targets is being able to look back at a list of things you have read.
If you’re not sure what to read, try this literature map which will give you a kind of moving mind map of authors based on who readers of an author you enjoy also read. Also have a look at the last Antonym, where I listed my top reads of 2023.
AI shadows
Check out this fascinating piece in Politico about people building bots trained on the writing of experts. You’d think the writing would copyrighted, but there’s a grey area here – like someone writing about things they’ve read, generative AI is summarising and synthesising. Like an automated version of the magazine The Week, I suppose. This article focuses on bots created to mimic relationship expert Ester Perel and psychologist Martin Seligman.
The creation of digital doubles of public figures has sparked a complex debate on ethics and legality. AI's ability to replicate individuals like psychologist Seligman and therapist Esther Perel with surprising accuracy presents a conundrum: the technology outpaces the legal framework meant to govern it.
While the potential for innovation is vast, the absence of clear legal guidelines leaves open questions of consent and intellectual property rights. New laws are being considered that would establish boundaries and conditions for the use of such AI-generated likenesses, aiming to protect individuals' rights without stifling technological progress.
Both Seligman and Perel eventually decided to accept the bots rather than challenge their existence. But if they’d wanted to shut down their digital replicas, it’s not clear they would have had a way to do it. Training AI on copyrighted works isn’t actually illegal. If the real Martin had wanted to block access to the fake one — a replica trained on his own thinking, using his own words, to produce all-new answers — it’s not clear he could have done anything about it.
For a less legally difficult example, you can get some leadership advice from Ulysses S Grant. I updated my USGrant AI bot recently by uploading the general’s complete memoirs (available copyright free) and it works surprisingly well.
It also offers sources for the advice, so you can read the bits of Grant’s memoir it is drawn from.
Predictions 2024
While we’re on the subject of my army of AI bots, don't forget you can ask the Zeitgeistbuster bot questions about your sector and predictions / trends for 2024 and it will scour fifty reports from banks, agencies, consultancies and think tanks to tell you what the clues are...
ChatGPT’s bot marketplace ( they call them GPTs) launches this week. It will be very interesting to see what people have built and what emerges from this new platform.
Democracy in danger
The Council on Foreign Relations (CFR), an influential US foreign policy think tank, published a report this week on AI's threat to democracy. The main concern is disinformation during elections:
In September, for example, a fake audio recording popped up on Facebook just two days before Slovakia’s elections; it used voice cloning to portray an interview between the leader of one of Slovakia’s political parties and a Slovakian journalist in which the progressive party leader appeared to be discussing how to rig the election.
And…
although these threats are not new, today’s generative AI capabilities will make these activities cheaper and more effective. Specifically, AI-enabled translation services, account creation tools, and data aggregation will allow bad actors to automate their processes and target individuals and organizations more precisely and at scale.
These are the emerging forces of generative AI:
Speeds up cognitive work – most things that require thinking can be don’t a little or a lot faster than before.
Assimilates knowledge or concepts from different sources. Systems like ChatGPT take their training set or data we give them and, guided by our prompts, create coherent, or at least plausible, syntheses from them.
The cost of content production moves toward zero. A bad actor could make fake videos or doctor photographs or write fake press releases before now, but now it’s getting cheaper and cheaper and easier and easier to do it.
You can see all three forces in play in electoral politics and we will see them play out in every other part of society, culture and commerce too.
Racial bias in AI
Sometimes the danger of disruption is that it will just make things worse. Just like previous technologies, the biases and inequalities present in systems can be amplified by AI, as McKinsey & Co’s article “The impact of generative AI on Black communities” [US-focused].
Without correcting long-standing patterns, gen AI has the potential to increase this racial wealth gap. Annual global wealth creation from gen AI is projected to be about $7 trillion, with almost $2 trillion of it expected to go to the United States, given its share of global GDP. US household wealth captures about 30 percent of US GDP, suggesting the United States could gain nearly $500 billion in household wealth from gen AI value creation. This increase would translate to an average of $3,400 in new wealth for each of the projected 143.4 million US households in 2045.
For more on the issues around racial bias and generative artificial intelligence, check out the fascinating ChatBlackGPT project by Erin Reddick.
Recommendations
Reading
Some snippets of things I’ve been reading this week.
The Future, by Naomi Alderman was one of my books of the year just finished. Tech billionaires try to dodge an apocalypse that they could have prevented.
The human race isn’t getting any better, we’re just getting different and more and faster and if you’re not getting better then more and faster is just the same as getting worse.
The MANIAC, by Benjamin Labatut is a fictionalised account of some of the 20th century’s greatest minds working on the edge of science.
Ehrenfest sought relentlessly what he called der springende Punkt, the leaping point, the heart of the matter, as for him deriving a result by logical means was never enough: "That is like dancing on one leg,” he would say, "when the essence lies in recognizing connections, meanings and associations in every direction." For Ehrenfest, true understanding was a full-body experience, something that involved your entire being, not just your mind or reason.
A God in Ruins, by Kate Atkinson, is a novel about a family that skips back forth in time and
They were all happy, this much at least he was sure of. Later on he realized it was never as simple as that. Happiness, like life itself, was as fragile as a bird’s heartbeat, as fleeting as the bluebells in the wood, but while it lasted, Fox Corner was an Arcadian dream.
And this passage sounds just like Barbara Kingsolver’s Demon Copperhead:
The future was a cage closing around him. Wasn’t life itself a great trap, its jaws waiting to snap?
Watching
I’m completely bowled over by Mr Inbetween, one of those series on Disney+ that is not on brand for Disney. About a violent Australian criminal whose oddly normal home-life and personal code create an odd, compelling ambivalence. Incredibly well made.
That’s all for this week
Thank you for reading. I hope there was something you found interesting.
If you liked it, nothing says “thank you” like a share!
Antony