This week: why AI is a lie; Apple finally reacts; an update on my writing; The Spike-related links; and recognising hope and joy don’t have to be corny with New Amsterdam.
ai_is_a_lie
This week, the UK government presented a new plan to make the most of AI’s “vast potential”.
As the fifth government to state their aim of becoming a “world leader” in AI, it’s concerning they don’t seem to understand it may already be close to its limit. And it is definitely not “intelligent”.
intelligence (noun)
the ability to learn, understand, and make judgments or have opinions that are based on reason
The Large Language Models (LLMs) like ChatGPT that are becoming so prevalent may be capable of learning, kind of, but they do not understand, nor can they have an opinion of their own. And they are entirely reliant on the quality of data used for training.
Defenders of AI will point out that humans are no different, but of course, we are. We can choose what data we decide is most relevant, what information may seem small but be inversely significant, what moral reasonings are applicable, and what conclusions to reach. Our results may vary between ourselves, but we are capable of explaining how we came to them.
This piece in Nature explains how LLMs lack essential metacognition for reliable reasoning.
Models consistently failed to recognize their knowledge limitations and provided confident answers even when correct options were absent.
Basically, an LLM can’t recognise when it doesn’t know something, will present an answer that cannot be correct as fact, and cannot explain how it came to that conclusion. The idea this could be used in any area, but particularly in medicine and treating patients, is scary.
The LLMs are like that kid at school who just repeated back everything you said to try to be amusing. There’s no independent thought or creativity there. It’s just a mimic.
I’m not going to rant now about the government’s decision to allow AI companies to ignore copyright laws and scrape whatever writing, images, video or music they want to train their models. I could go on for pages. Many others have already laid out their criticisms, like here and here.
I’m still hoping the government will be convinced to change their mind, if they can look past the promise of vast sums from the tech giants. But I will say this:
The idea that creatives should need to opt out of having their work stolen is like a shop owner needing to tell everyone who walks in they are not allowed to shoplift, otherwise anyone can take whatever they like without punishment.
I’ve signed a statement against AI training, against the flagrant disregard of creatives and copyright rules, here. If you’re a creative of any kind, please consider doing the same.
apple_finally_reacts
Having just written about the lie of AI, I return to the subject of AI making up and reporting lies, although the irony isn’t lost on me that this is about them (finally) being prevented.
Last week was the second time I’d covered Apple’s “Intelligence” collating news stories from other sources into incorrect summaries, sometimes completely contradicting the original report.
Apple had been ignoring calls to take this function offline until they fixed the issue, but have now suspended it for News and Entertainment summaries. It will italicize text on the lock screen to make it easier to separate summaries from notifications, and will warn users in the Settings about potential errors and make it clearer the function is still in beta (as per The Register).
It’s a start but I’m not sure this is sufficient, and will be watching carefully when this is rereleased.
progress_report
As anticipated, my day job is encroaching on my time and will to write in the evenings at the moment, but I did make some progress. The chapters I’m working on for part 3 are particularly interesting and fun to write, and quite different from those in part 2.
I’m pleased that, despite all three parts in the volume sharing the same theme, they all have a different feel. It’s made writing it more engaging, and I hope will make it more entertaining to read.
status.vol2
The Spike Volume 2 will contain three separate books from the perspective of seven characters.
Part 1 – draft 2 complete; further chapters to add.
Part 2 – draft 2 complete!
Part 3 – draft 2 written up to chapter 20; forty-six chapters to go.
The intention is to complete draft 2 of part 3 by the end of February.
Part 1 is currently much shorter than the others, and I want them to be closer in length, so I will need to decide how to tighten parts 2 and 3 slightly, and add more to part 1 – I have some exciting ideas to expand it.
Then a readthrough of all parts will determine how much revision is required.
The ultimate aim is to have Volume 2 finished by the end of 2025 for publication early 2026.
connecting_links
The Spike is set in our world, incorporating real events; the links below are relevant to the themes and overarching storyline, and may or may not provide clues to the direction of the series.
I do not necessarily agree with or endorse any of the views within.
Nvidia: fully autonomous cars ‘not close’ to production reality
China wants to build massive solar station in space — it’s like a ‘Three Gorges dam’ in orbit
Meta’s MAGA Metamorphosis: Project 2025’s Digital Revolution Takes Shape
The Shadow of Spyware: UN Security Council Grapples with a Global Threat
Safer, Stronger, Smarter: Scientists Develop Game-Changing Quasi-Solid Battery
weekly_inspiration
Every week I share something that’s inspired my creativity.
This week, I let the TV show New Amsterdam heal me and put my anger of the previous week aside.
if you don’t know, New Amsterdam is an American drama series set in a public hospital. Whilst there are of course sad moments (inevitable and necessary when dealing with mortality and humanity), the general theme is of hope and of trying to make things better for everyone.
Much popular entertainment nowadays tries to be cool or gritty, presumably due to negativity bias, and that may be why I haven’t heard many people talking about New Amsterdam. I’m a cynic myself (in case you hadn’t noticed) so this isn’t the kind of thing I would normally watch, but it’s so well done, it’s hard not to enjoy it.
The writing in particular is excellent, with subplots that feed into the greater arc of each episode, sharp dialogue, and some genuinely emotional moments.
It’s good to remember it is possible to write positively without it sounding cheesy or schmaltzy.
Seasons 1-4 are available to stream on Amazon Prime in the UK.
https://www.imdb.com/title/tt7817340/?ref_=nv_sr_srsg_1_tt_7_nm_0_in_0_q_new%2520a
What’s inspired you this week? Please share in the comments.
See you next week.
Arrowmaker Books 