Wednesday 31 May 2023

Challenges for AI: if you're so bloody smart ...

It is suggested that the acid test for AI - indeed, for it to qualify as AGI - is that it should start generating New Science: hitherto unknown but recognisably scientific advances that can be tested against the usual criteria for such things.  We read that 'ordinary' 2023-vintage AI has been used in developing a wholly new antibiotic for use against Acinetobacter baumannii, a nasty strain of superbug.  OK, and all power to the team that achieved this: but it's not quite an AI achievement per se: we're still in "useful software" territory.  Couldn't have been done without software?  Well, neither could much of what I've been doing today by way of working from home.

So - I'd like to propose a test.  To my mind, one of the weirdest "non developments" of the past 100 years has been the complete failure of bio-chemical science to come up with artificial blood.  Surely, that would be up there with genuinely efficient large-scale storage of electricity** - crack that, and you can have your university re-named after yourself.  "Being a big help with finding antibiotics" is great.  Devising artificial blood - there's the real test.  Go for it, Mr A(G)I.

Any other tests that readers would like to suggest?



** Yes, of course I know that it's never really electricity that gets stored, it's electricity-into-something-else-&-back-into-electricity-again-later. 

Thursday 25 May 2023

Rolling euphemisms, wokery, & the glories of the English language

Keeping up with whatever is the approved euphemism has always been a great lark.  We read that "obesity" now needs to be replaced - and not with "too fat", either.  The Beeb hasn't caught up with this one yet - you do really need to be on your toes.  Fortunately, the English language is replete with a wide range of colourful alternatives.

It was ever thus: whole histories could be written on the rolling evolution of the approved terms for, e.g. death, various disabilities, and descriptions of ethnicity.  Up to a point, this is to do with kindness: when a particular conventional usage becomes a term of abuse or mockery, it's time to coin something new.  Which is fair enough, albeit we sometimes get empty phrases like "learning difficulties" which covers so broad a spectrum of phenomena as to be rendered meaningless.  I have learning difficulties when I keep being interrupted, but I doubt I qualify for charitable largesse.

There's a nastier dynamic at work, though.  When anything is being politicised by the identitarian left, terminology becomes paramount, the left being beset by over-intellectualisation, theorising - and faddery, one-upmanship and ideological purity.  And nothing is more enjoyable than witch-hunting.  So nothing is neater than to keep changing the rules by making some term in common usage a shibboleth: utter it, or say it wrong - and you're damned.  Hurrah! - another heretic exposed and harried to the point of despair!

And this need to "be on your toes" ensures that only the elite of lefties at radical academic institutions can ever be up to date in these matters.  Everyone else is guaranteed to be behind the curve, beyond the pale, and in need of re-education and groveling self-flagellation, at the very least.  But maybe total cancelation and ruination.  Oh, the horrors of feeling oneself in thrall to such vicious little shits - who increasingly call the shot, not only in colleges, but in HR departments across the English-speaking world, depressing though it is to acknowledge.

If there is hope, it lies with the proles, as Orwell famously wrote.  They'll have some ideas on the subject of obesity, and one or two other things besides.  Who ate all the pies, eh?


Tuesday 23 May 2023

After Bakhmut: let the trolling commence

For all intents and purposes the town of Bakhmut has fallen, albeit that Ukraine has taken the opportunity afforded by the exhaustion and tactical ineptness of the invading forces, to pinch it on either flank.  For the time being there will be no Russian advance on Slovyansk and Kramatorsk, which were supposed to be the next towns down the road, last summer while Putin still enjoyed a (tiny) bit of momentum.  Informed opinion varies as to whether the 4-month street-fighting defence of Bakhmut was as advantageous to Ukraine in ratio of casualties as merited holding on for quite that long.  It certainly blunted the Russian Donbas offensive to the point of total standstill, depriving Putin of one of his stated main aims.  Further, it fixed Russian forces to an extent they would not have welcomed.  My own observation would be that Ukraine has not thus far demonstrated gratuitously bad military judgement, or overly emotional attachment to any particular piece of real estate.  (Most mistakes have, understandably, been ones of caution - always easier to identify in retrospect.)

We now await the long-heralded Ukrainian 2023 counteroffensive.  As I endlessly say, there's no such thing as strategic surprise ... so it will be interesting to see what tactical surprises they have up their sleeves.  It's imperative they have something clever in mind; otherwise, with no prospect of air superiority they are potentially in trouble against a lengthily-prepared, albeit unsubtle and static defensive plan, established in depth.  Frankly, there are no good 20th century precedents that I know of, notwithstanding that static defences have often been swept aside - with air support.  The major points in Ukraine's favour are:

  • excellent intelligence; the initiative; and many months of active planning
  • the front is huge, and Russian forces (of poor fighting quality on average) are thinly spread.  Their defensive depth looks excellent on the map, but only a small % is actively manned
  • defence in such circumstances relies upon good communications, coordination and decision-making under conditions of chaos, plus adroit deployment of highly mobile reserves: and the Russians have thus far been found badly wanting in all those regards
  • I'm not convinced Putin dares hazard his airforce, the only wholly irreplaceable asset he has
I shan't bother to list Ukraine's difficulties.  Anyhow, we may not need to speculate for long.

*   *   *   *   *

I was looking back at an excellent June 2022 article by Adam Tooze the other day, and was amused to take a peek BTL.  We've been honoured with some pretty silly trolling on C@W back in the early months of this war - who knew we merited the attention of the Russian troll factories? - but Tooze got both barrels.  Samples: 

... the [Russian] breakthrough at Popasna** and the developing envelopment of Slavyansk, as such defences tend to hold in place until they break, and then they break quite quickly. In this war the Ukrainian side has been the rigid and inflexible one (dig in, defend anything to the point of encirclement and annihilation), while the Russians moved and maneuvered over a 800km length of battlefield, changing axis of attack and intensity many times. The results are slowly becoming visible - the Ukrainians are losing the war 

Russia’s outnumbered, highly mechanized force has constantly set the rules of the engagement and Ukraine’s massive mobilized force, it seems, is fundamentally reactive.

it is very obvious that Ukraine is retreating everywhere. ... a full collapse (of Ukr) is coming

the Russians are winning in a very loss-effective manner. Once beyond the in-depth defences of the Donbass, and with much of the professional Ukrainian army under the ground, a prisoner of war, or injured beyond recovery to the field (or refusing to fight as has been shown in multiple cases), the war tempo will move much faster. Zelensky and his government have made a major strategic error in pouring in men and arms into a salient thats will inevitably collapse, a meat grinder being used by the Russians to destroy the Ukrainian army ... Zelensky throws more and more of his army into this cauldron. Once the cauldron, and much of the Ukrainian Army, is finished who will be left to defend the rest of Ukraine?

The Ukraine military has been defeated for two months already. I don't know what kind of funding you are smoking, Adam, but the only successes in Ukraine are the people cleaning up all the dead Ukrainian soldiers. It is beyond disgusting that people are pretending that those facts are not real ... They all should have surrendered long ago and every death is just another mark of American hubris.

Admit it. Russia has won, will win, will demand all sanctions removed. And Ukraine, for its efforts better remove crooked Zelenskyy and admit that their sacrifices were in vain. 

Hmm.  I think we shall see more like this with which to decorate Ukraine's counteroffensive.



** Russian "breakthrough"?  Very short-lived; subsequently routed; and Slavyansk never troubled. 

Wednesday 17 May 2023

AI applications in the military

Recently, under a post about the fate of truth in 2023, I included (BTL) a link to a bot-written article on the issues around the use of AI in the military; and commenter Jim opined: 

Most probably serious development is going on to specialise in legal argument, military strategy, economic planning.

I said we'd come back to military applications, which do indeed proceed apace.  Many in the military (in many countries) hope for AI and other computer-based solutions to a whole range of challenges - not least because they think they'll be a lot cheaper, in both blood and treasure.  Of course, from time to time the 'smart', asset-lite approach comes badly unstuck at the hands of determined infantrymen, as Ukraine has taught the lazy Putin and his attempt at a 'smart' lightning campaign last year.  (He seriously fell for the plan to persuade Ukrainian officers to surrender by texting them all a threatening message on their personal 'phones, FFS.)

That said, just as fast-learning computers will cheerfully thrash anyone at chess nowadays, they will also thrash any fighter pilot in simulated dogfights.  Yup, they can get right inside any human decision-loop, and destroy the mortal sluggard with remorseless efficiency.  Who doesn't want drones like that?

But of course, issues arise about less-than-perfect "decision making" even by these computerised paragons.  How are they to be employed?  One answer is: in some applications, perfection never has been expected, and isn't necessary - just a very good probability of turning in a very good result in very quick time.  The rest is just, well, fortunes of war - casualties & all.  Always has been.  

Here's an example from my (somewhat dated, but recently updated) personal experience.  It's trivial enough in its own way, and won't surprise anyone: but it carries some lessons.  Starting around the time of the First Gulf War, a phenomenon that was long anticipated became real: total and utter data overload in the sphere of aerial imagery (inter alia).  More imagery was being produced, of areas of very high significance (e.g., known to harbour Scud missile launchers), than could be fully processed by the available highly-skilled analytic personnel.  Enter software that could run its unblinking eye over the coverage and identify + highlight straight lines, rectangles and circles that are of a size likely to be of interest.  There aren't too many of them occurring in nature: they'll almost always be man-made.  The human could then zoom in on what the computer had found, and discard those that were false positives.  Had the computer missed a few?  Maybe - but so what: they were never going to be found anyway!  And the next version of the software would be better.

To make a long story short, many learning-cycles later, computers nowadays have been trained to identify from aerial imagery, and put a name to - with high success rates - specific types of equipment: any ship or aircraft whatsoever, and quite a lot of vehicles.  Does camouflage still work?  Of course it does, sometimes.  Are there sometimes errors?  Yup (and the humans are quite prone to this, too.)  But 95% accuracy (say), carried out very fast, is a very acceptable result: like Patton said: a good plan executed violently now, is better than a perfect plan next week.  (In due course, 96% will be better ... they're learning all the time.) 

That's just one micro application in the military.  The point is: perfection is not required.  In the fog of war, something very good and very fast is highly desirable.  AI is going to provide a lot of that.


Friday 12 May 2023

100% Mortgage? In a banking crisis?

Actually, the new Skipton product - 100% mortgage, no deposit, but (they say) stringent borrower qualifications required (primarily, a good prior rental payment record), plus a small interest-rate premium - all seems sound from the lender's risk-management point of view.  One assumes they'll be pretty careful about getting their own valuation done, too.  Structurally this is all good, pragmatic RM - and moderately creative, to boot.  

When Miss D got her first mortgage recently she was in pocket immediately (on a current account basis), the payments being less than she was previously paying in rent.  That's a dynamic the Skipton product arbitrages neatly.  Don't know if that situation prevails everywhere; but it does in her part of London.

Commentators seemed to be fixating on the potential for negative equity.  Well, yes - but both parties ought to be able to take an intelligent view on that.  On the lender's side, one assumes more pragmatic RM, relating to the specifics of the property and the profile of the borrower.  And for the borrower: hey, did you want (a) to get into property, (b) & without putting in equity - or not?

Go for it, Skipton.  Nice structure: hope you've got the details right.


Wednesday 10 May 2023

France: always hanging back, always wanting to share the credit

A first today - a French Air Force recce aircraft over the Black Sea!  There have been non-stop US, UK, Italian, even Polish, Swedish and Finnish recce flights in support of Ukraine, but this is the first (openly detectable) French one.

Now, given that they have had the capability all along, why would this be?  It's the usual French thing.  Not us, gov, we're not like those warmongering Yanks and Brits, we're not feeding this conflict by providing one side with intelligence (actually, we don't have any), we're not helpfully wargaming strategies with one of the protagonists: we're responsible peace-loving people, neutral, here to negotiate like statesmen!  We'll talk Putin round, we'll get Xi onside ...

Oh, what's that? - the big Ukrainian counter-offensive push is on for next week?  OK, we're in, we're in!  They couldn't do it without us!  A seat at the Top Table for us!

Running with the hares, hunting with the hounds, or what?   Or perhaps the playground coward, rushing in to kick the guy who's already down on the floor.  Ho hum, better late than never.


Friday 5 May 2023

'Where is truth' in 2023?

'What is truth?' said jesting Pilate; and stayed not for an answer
A commonplace assertion is that we live in a 'post-truth' world, centuries after the Enlightenment supposedly (under the Whig theory of history) set mankind on an irrevocable path towards, at the very least, valuing and seeking after truth.  That's Truth as it would be understood by a philosophical Realist: existing objectively and independently of anyone's knowing it, or even having the capacity to know it.

Even within this worldview everyone has long resigned themselves to co-existing with all manner of deflections from the straight and narrow: the dogmas of religious fundamentalists (to be tolerated in a worldly manner, up until the point where they start to impinge on everyone else); an intelligent degree of relativism (provided Realism is not ultimately ditched); acceptance that politicians of all sorts are doomed to deal often in half-truths and evasions; knowledge that science only proceeds from one approximation to another, better approximation; that the capacity of most people for understanding and/or dealing with all the truths that are out there to be understood is distinctly limited, with various strategies being needed to accommodate this fact.  

At the end of the day, there's the smug intellectually comforting Realist consideration that a truth ignored or denied is still out there, and will potentially break your toe if you kick out against it.

All that said, the advent of Blair and Bush Jnr ushered in an era when even western politicians, supposedly 'better' than the Putins of this world, increasingly cared not a fig for truth, but merely and shamelessly called for Alastair Campbell to come up with one of his sophisticated utter-bullshit operations to suit the needs of the hour.

The Common Man has always had an interesting role to play in all this.  Firstly, he's not very concerned with the truth or otherwise of the dogmas swirling around (which he knows are well above his pay-grade); he just wants to know what he has to repeat solemnly in public, as and when required, in order for the (current) authorities not to be after him for unintended heresy.  If the dogma changes, he just wants to be told what are the new shibboleths.  In times of religious oppression this becomes an acute and worrying business, but fainter, less lethal manifestations are always in the air: as Homer Simpson says, the highest wisdom is - never be the only person in the room that's laughing.

Secondly, though, the Common Man does know some 'hard' truths very well indeed - most significantly, those that impinge on his making a living: these are the truths that will break your toe.  If he's a blacksmith, he's never going to 'believe' that you can hammer cold metal into a horseshoe, even if some idiot Inquisitioner forces him on pain of death to assert it in the middle of the town square.  Of course the average Inquisition, not wanting to be made mockery of, will stick to things that can confidently be asserted in the certainty that no hard-to-ignore practical disproof will ever be at hand (transubstantiation; virgins on offer for martyrs, etc etc).  Even so, Stalin's and Mao's dogmatic embrace of some outright lunacies in the scientific sphere forced any number of Soviet / Chinese engineers etc to endorse them publicly but ignore them without comment in their daily activity:  these ones require active doublethink.  (It has always seemed likely to me that the Stalins of this world don't much care about this phenomenon, just so long as everyone lines up for the 'loyalty test' of repeating faithfully the dogma of the day.  In fact, for loyalty-test purposes alone, the dafter it is, the better.  And the Common Man is generally OK with doublethink.)

*   *   *   *   *

What has caused me to muse on all this in May 2023?  The other day I watched Official Secrets, a rather good, fact-based movie (haha!) about the GCHQ whistle-blower who leaked a document in which the Bush operation sought the UK's assistance in blackmailing various UN delegations in order to swing a UN vote that would formally legitimise Gulf War 2 - not that neither Bush nor Blair ultimately much cared about that legitimation.  Whistle-blowers, of course, are just that small minority of people that can't bring themselves to do the doublethink.  It's a small minority indeed.

But ancient history be blowed: look at what we are faced with right now (- you can pitch in with your own favourites):
  • Starmer systematically and merrily reneging on every one of his "10 pledges", with such pious outriders as David Lammy and Polly Toynbee all rushing to assert this doesn't matter a whit
  • the complete absence of any reliable accounts on significant matters of the day such as the blowing-up of the Nord Stream pipeline
  • the complete works of Donald Trump, Boris Johnson, Megan Markle etc ad nauseam
A particular favourite of mine - the 'progressive' line on trans self-identification - is thrown nicely into relief by this story in today's Graun (a paper which is actually, and to its credit, trying to drag itself painfully away from the progressive line).  This piece is really worth reading - it is absolutely hilarious in its straight-faced delivery of utter nonsense. 
NHS treatment algorithms ‘not taking transgender patients into account’:   Medics say trans people being put at risk by lack of evidence on how to assess them by gender-based metrics
Or to put the matter more simply: as 99.9% of humanity knows very well, "Trans women are, errr, men"  Now that's the kind of truth that really breaks toes when kicked against - ask Nicola Sturgeon.  

So: God Save The King!  And don't press me (or the Common Man) too closely on that Holy Anointing Oil ...