somewhereoutth

Yet another [1] AI winter [2]

[1] https://en.wikipedia.org/wiki/AI_winter#The_setbacks_of_1974

[2] https://en.wikipedia.org/wiki/AI_winter#AI_winter_of_the_199...

However 'normal people' buying into the stock market via 401ks or otherwise likely (and arguably sensibly) will be in index funds, that of course are exposed to the bubble via (grossly?) inflated tech stocks. Effectively their current pension/savings contributions are being clipped by whatever the delta is between now and post bubble. Time in the market and all that, but still it might be a hefty haircut.

LLMs: The chemical weapons of public discourse.

The cleanup is going to be a grim task.

Although INTERCAL's "come from" provided the opportunity to implement a gloriously concise multi-threading syntax/mechanism [1]!

[1] https://esolangs.org/wiki/Threaded_INTERCAL

1. Understand that code that has been wholly or partly LLM generated is tainted - it has (in at least some part) been created neither by humans nor by a deterministic, verifiable, process. Any representations to its quality are therefore void.

2. Ban tainted code.

Consider code that (in the old days) had been copy pasted from elsewhere. Is that any better than LLM generated code? Why yes - to make it work a human had to comb through it, tweaking as necessary, and if they did not then stylistic cues make the copy pasta quite evident. LLMs effectively originate and disguise copy pasta (including mimicking house styles), making it harder/impossible to validate the code without stepping through every single statement. The process can no longer be validated, so the output has to be. Which does not scale.

Is this about freedom to vs freedom from?

The first freedom of course is freedom from fear (e.g. from fear of being snatched from the street just because you are brown and speak with an accent)

Ah indeed - actually 2 years ago. I believe 21.2 (Victoria) was released in 2023:

https://tuxcare.com/blog/is-linux-mint-based-on-ubuntu/

Doesn't time fly!

Ah I think what is happening is that a brand new laptop may not have had its drivers added to the latest download/image version of your chosen distro. Hence the need to connect and update to get those drivers.

For me, it was a ThinkPad X1 Gen 11 and Mint 21.2 (MATE) about a year ago.

Wifi never seems to work 'out the box', so be ready to connect to the internet using ethernet to download the latest drivers and thus get wifi to work.

Interestingly, in the Lambda Calculus, where everything is a function, a standard representation for a natural number n (i.e. a whole number >= 0), is indeed a function that 'iterates' (strictly, folds/recurses) n times.

E.g. 3:

(f, x) => f(f(f(x)))

My understanding is that the 'extra thing' is control flow - blocks can force a return in their calling scope. For example a loop that calls a block may be terminated/skipped by a break/continue statement in the block itself. However I'm not a Ruby programmer, so please check my working.

> the claims that it's not useful

There are many credible claims that not only is it not useful, but that it is actually causing serious damage.

> The development of AI will clearly have society-changing effects, just as the arrival of the internet and mobile communication did.

Or, maybe the economics can't be made to pencil out, and AI is abandoned (again).

> The impossibility lies in knowing the speed of adoption, and which lumps of capital will earn extraordinary returns and which will end up getting torched.

Almost certainly all of it will end up getting torched.

I've often idly conjectured that a 'perfect virus' would quietly replicate, and then occasionally flip a one or zero here and there. The damage would be imperceptible to begin with, but then eventually critical 'ones and zeros' would be affected, and it would be too late to try to figure out what systems were clean.

Never in my wildest imaginings did I come up with something as diabolical as LLM generated code.

If I understand correctly, the cost of training each next model is the biggest cost - and much of why so much capex has been required. I have also seen it stated that e.g. OpenAI could be profitable right now if they stopped spending on the next model (not sure I'm convinced of that though).

Given that a half (or 80% or whatever) trained model is worthless, there is a threshold for each new model cycle to continue. As soon as the anticipated ROI fails to justify that cost, new model production stops. Dead.

Old models will still be used, but will become more and more out of date, and less and less useful. This will be the end of AI.

(sorry, paraphrasing/developing my comment on an earlier thread)

My understanding is that the cost of training each next model is very very large, and a half trained model is worthless.

Thus when it is realised that this investment cannot produce the necessary returns, there will simply be no next model. People will continue using the old models, but they will become more and more out of date, and less and less useful, until they are not much more than historical artifacts.

My point is that the threshold for continuing this process (new models) is very big (getting bigger each time?), so the 'pop' will be a step function to zero.

Assuming you are living in a modern economy - including the US - around half of it is owned/run by the government, i.e. half of it is socialist (social ownership of the means of production)

One of the pressures causing migration is global warming screwing up agriculture, particularly in poorer regions.

> Give an LLM a Jira ticket and a codebase and it can push a PR

Not really though can it, unless that codebase and Jira happens to pattern match correctly somewhere.

I'd say that the difference is just that LLMs have a natural language interface (for good or for ill).

But we didn't do this with spreadsheets or word processors - and those tools really did replace small armies of clerical and secretarial workers.

> developer somewhere between junior and mid-level

Why the insistence on anthropomorphizing what is just a tool? It has no agency, does not 'think' in any meaningful manner, it is just pattern matching on a vast corpus of training data. That's not to say it can't be very useful - as you seem to have found - but it is still just a tool.

The highest body of the EU is the Council. Nothing happens without the approval of the Council. In comparison, the Commission is merely the civil service or secretariat, answering to the Council.

Each member state has a seat at the Council, and for almost all issues a veto. Each member state is democratic, therefore the EU itself is entirely democratic. That doesn't of course mean the right decisions are always made!

One extra 'thing' to need - at the moment I know that I can play music through anything that has a line-in, with just a cable. However Bluetooth seems to work ok - for devices that support it.

In modern economies, banks create money by originating loans. The government controls this process with various policy levers.

Samsung Galaxy A40 checking in.

It's small, has dual sim card sockets, and a headphone jack.

I'm not sure how I'd replace it to be honest.

The EU Council is the highest body in the EU (not the Parliament, especially not the Commission - who are basically the civil service or secretariat for the EU).

The EU is founded on the pooled sovereignty of the member states (unlike in the US, where the reverse is the case). The Council represents those member states (each has a seat), and so holds this pooled sovereignty.

Completely not true and in fact dangerously (willfully?) misleading.

Housing is much more complicated than supply and demand - for a start houses are not fungible. Local conditions are extremely important.

A lot of high house prices is due to financial engineering - for example long mortgage periods at low interest rates. Just one of many mechanisms to make the rich richer by making the poor poorer. Any momentary affordability benefit is soaked up by higher prices, and those mortgages have to be paid off for longer with much higher total payments.

High house prices are also caused by demand elasticity - things like AirBnB, 2nd, 3rd, 4th homes for the super rich, and housing as a place to park funny money (regardless of whether the house is actually being used to home people) mean that normal people can't compete with this non-home demand.

We could go on.

> you build a world model

The foundational conceit (if you will) of LLMs is that they build a semantic (world) model to 'make sense' of their training. However it is much more likely that they are simply building a syntactic model in response to the training. As far as I know there is no evidence of a semantic model emerging.

In the UK we have, not an 'owning a property tax', but instead a 'living in a property tax' (Council Tax). At one point there was an attempt to impose a 'living/breathing tax' (Poll Tax) but even we couldn't stomach that, and rioted until it was got rid of.

“How could they see anything but the shadows if they were never allowed to move their heads?” Plato, The Allegory of the Cave