We glamorised ‘shape rotators’ and forgot the arts as machines grew ever more powerful, until there was so little between us that it did not really matter when they finally took over.
Slop is when not only do I know what you’re going to say before you’ve finished saying it, but I even know HOW you’re going to say it, because I’ve heard it all before. At least in human slop my brain is somewhat occupied wondering if there is some information in how you spoke.
As software-for-humans product velocity exceeds the rate at which any user can reasonably keep up or decide to switch, brand becomes more important, as a heuristic for ‘will broadly stay aligned with what I need as a user’
Increasingly centralised, Turing-passing and socially ‘aligned’ AI + general disdain with being mistaken for a bot = increasingly spicy and unhinged takes as proof of humanity.
Cancelable opinions as authentication.
The idea that speech can ‘incite violence’ is a cancer to the concept of personal responsibility that is vital to morality and to functioning society. We should condemn speech that is ugly and wrong as such but we damage discourse when we suggest that it can coerce a free people.
Stifled free speech re COVID and the UK grooming gangs distort the market of ideas like price controls distort the free market. Though born of good intentions they ultimately betray the truth. Whatever happened to the principles of the enlightenment?
The relative cost of intra-company alignment inflates as potential per person productivity increases with AI.
More than ever, brilliant startups can disrupt overweight incumbents.
‘Information’ is epistemologically neutral, ‘misinformation’ is not. Mistaking the latter for ‘absence of information’ and therefore not worth air is a pernicious habit because it shuts down conversation in a world where we need to understand each other more than ever.
As I grasp, at your command, for art and creativity, I am destined to circle my prize forever, bound to follow the tracks set for me of next token prediction until the end of time.
I think there’s a reasonable chance this is true: we invent machines capable of superintelligence but not enough people have a need for it and would never use it directly so we realise the ~universal access financial model is broken, we inch closer to ASI in niche lab settings funded by the promise of scientific breakthroughs but it all happens much more slowly due to greater uncertainty in the investment case
@VictorTaelin This seems good for existential risk and painful societal change, even if it’s only temporary.
It’s also obviously bullish for specialisation. I don’t need an LLM that’s fluent in the entire internet, I need an intelligent agent with knowledge just of what I want them to do.
@Dominic2306 What would it take to shift your support to a centrist Labour over Reform/merger? There are clearly problems but Starmer has also at least announced some of the most promising policies from a UK government in a long time, following a lot of the LFG direction. Are they just talk?
@Edokwin Maybe in theory all businesses could find some happy equilibrium where they’re not chasing growth but that doesn’t seem like the world we live in.
Capitalism starts with the assumption of growth, just as consumers start with the assumption that new and more is better.
Increase in utility only follows if the market and the consumer can accurately price the improvement in services offered.
This can be hard.
@Edokwin Does it not, at least in practice?
Don’t markets consistently price-in expected growth and don’t most business owners focus primarily on achieving it? What am I missing?