Blog

Archeo-ML

At a tangent to many great posts on, e.g., LinkedIn, regarding kuh-rypto-stuff mostly and how that should be done, e.g., by some expert writers and others, two things sprang up that would make for excellent ‘AI for Good’ study projects:

  • Quipu deciphering – typically, cryptographers haven’t been able to tackle the decryption but wouldn’t machine learning, possibly in combination with advanced language processing analytics / methodology, be able to establish non-human-recognisable patterns ..?
  • Celtic ruin runes to real conversation – a typical interpolation effort, it seems, squared with pattern matching errr… identification first. Oh, and where I wrote Celtic ruin runes, it should of course read ‘old French jumble’, easing the challenge considerably but would have been semi-pleonastic.

Yes I was looking for a 3rd challenge, ’cause I dislike devil’s numbers since they refer to duality that is such Luciferian cause for Disruption in all the world’s wisdom traditions. Even the Chinese Yin-Yang has this ‘need’ for balance, and the Native Americans / First Nations have the Trickster being the unruly against the Stable One [noting that there, too, the Trickster is fully 100% essential for the Creation and onward creations to exist at all ..!], and … , and …:

And where are all the students that want to get all the attention that solving an above riddle may nay will bring, globally ..? Please report to your local grad advisor; I’ll be glad to support your efforts with praise and attention.

And this:

[Not a riddle as much … Casa de Musica Porto, detail with attempt at art’sy effect]

Missioning your visioning

Just a repost; hardly anything but the footnotes had to change from that post from over two years ago..:

I was triggered recently[1] about some very common mix-up, and suddenly saw a spark of insight. Since the seeing concerned me, it was at a distance of course (not), so I’ll share it with you to see whether you recognise the somewhat local phenomenon and its ramifications and cures.

This all being about the top-level structure for any organisation, i.e., vision, mission, strategy.

So often mixed up qua priorities and order.
Where a great many do think one starts with the mission, thinking endlessly wrongly that one determines the mission first, then … oh vision how can we define that, then, to cover the actual second step, and strategy oh we will not be able to define that.
Because of the mix-up you can’t. Because vision comes first. First define what and how you think the world will look like in somesomewhat more distant future. ONLY THEN [skipping the boldface] one defines whether and where one’s (organisation’s) place is thought to be within that; mission. Only then does one define how to get from the paltry Today’s position, to where the bright future is, and which path to follow; strategy.
When you begin with the mission, you utterly falsely assume there’s a place for you in the first place which, when you think that way, there is ever more certainly not.
When you start with the vision (which can be grand, being without you in the picture will help to that end…), you would also need to think about what your raison d’être in that world view would be. You have no right to existence [note that we’re discussing organisational rights, not natural persons’ rights for which the opposite does hold; ed.] other than making the future better – than any of your competitors can. If there’s competitors that can, in principle given their today’s capabilities, in fact do better than you, you’ll have to find something else to do as the world, by Pareto’s comparative advantage reasoning [which is one of the few economists’ reasonings of any value at all when without the ‘ceteris paribus’ totalitarian destructive lies; ed.], is better off when you leave it to others – the world better off being your purpose or you have no place on this planet. Anything related to profitability has no, zero, rien du tout, nul, nada, place in a mission statement since it’s a derivative requirement (sic) only, towards some specific but not therefore important, stakeholders known as shareholders or investors that, if your mission is true and virtuous, should be utterly grateful for the opportunity to be allowed to invest in your strategy and would even have to pay rent, not receive it. For organisations, it is only a requirement in as far as reserves are built, for hard times and even there one could insert some evolutionary theory that sometimes extremely improving mutations are wiped out by some unfortunate accident (act of nature) before becoming a species-wide improvement — same, for organisations that are set to become most beneficial to the world already on the road towards fulfilling their mission but fail halfway due to adverse but insufficiently bendable market conditions. Bad luck. Move over. Machiavelli’s Fortuna again. Read his Original to see how sobering he meant that and that you can be (Aristotelian-)virtuous all the way through.

So, first vision, then mission. And only then, strategy of course. Being the course. You need to take; roughly. Bending and shaping as you go along; sailing to an up-wind buoy. Setting the boundary conditions – including the boundaries of means you’ll allow as not all means are allowed for just any ends, no end may be enough for some means, most ends are not worth pursuing in the first place. You get the drift; if the means are not ethical, virtuous, the ends will never be. Excepting only a handful of situations like war/genocide, large-scale natural disaster, et al. where one may have to shoot the bad guys to prevent them shooting the good guys.[2]

But hey, this was just the spark. About the wrong-order its cause(s). The rest … good for PowerPoint presentations on the subject I’d say.
[Edited to add: Next try to understand the utter ridiculousness of the Dutch ‘beleid’ as translation of ‘policy’. Meaning (in both languages, creepingly) not what it has meant over the past decades, centuries, but more like ‘petty micromanagement ruleset’. No more general picket post placement in the distance but rails, shackles…]
Leaving you with:

[The lands you defends, into the fuzziness; (from) Salzburg Castle]

[1] Yes, in a conversation to establish whether I’d want to work somewhere, and they’d have me. Both ways, the result was Unsure But Let’s Keep On Talking. They had a Mission on their glass door, and a Vision, and Values – you understand that Mission and Vision were a jumble of Calimero-aspirational buzz, and Values were a big fat Mehhh. If you have to post them on your door in the hall to remind any employee coming to work day in, day out, you’re .. well, not quite living them, are you ..?
[2] Tell-tale, this [1] club had Values there, not Strategy. The latter, didn’t even surface in a second conversation at all. As if one bumped into a fluffy ceiling when trying to raise the level, before being pulled back to mundane hire-warm-body work descriptions. Some sparks of want to move forward and some slight claim of record there but hardly any self-volunteered methodology hints or so…

All are Us

Lately, there have been umpteen blog posts about the demise of a great power. Like, here and my own here [ w/ links].
Was reminded that in the core, the underunderunderlying problem had already been caught in the Classic. Not plural: This one. If, big big if, you can see. And a long way from the good [???] old past in this which turned out to uplift but not be realised…
Maybe there’s hope in this Atlantic piece.

On which side are you?

Anyone knows where the artist(s) have gone ..?
Anyway; study!
Yes it’s about:

[Just screenshot, but it captures the full lyrics quite well]

ICYMI; Risk Management Requires Quite Some Wider Perspective

Yeah, I’ve blogged about risk management quite a lot, lately. The most of content was about RM1 versus RM2; the former being the day-to-day control over negative events through controls, so as to get to the Objectives set by … well, that follows below, and the latter, being about tornado charts on fishboning all-than-needs-to-go-well-to-make-this-a-success factors, qua risk distributions.

Also, about the ubiquitous (total) proof that heat maps and similar semi-quantitative RM1 is tort, pure fraud. No need to sugarcoat it, right? The text “All models are wrong, but some are useful” was a joke but dunces took it as an alibi to still use their crappy models. Which is plain stupid, whatever ‘arguments’ one might bring against that qualification… Plus, one’s not dealing with simple things [or must be relegated to child’s play work and pay..!] but with complexity. As here. But then, as said I’ve blogged about that sufficiently too, as here. Model risk stands at 100% #fail.

Now, suddenly I found this piece. It re-iterates the above RM1 ↔ RM2 distinction, and adds quite some interesting angles [what follows is my take from it, with some personal sauce extension…]:

  • Every objective is (part of) a representation of some ultimate goal(s). Without clear goals, no suitable objectives, raising the risk (sic) that when objectives might be achieved, the goals still aren’t;
  • Or rather, the goal (post)s may have moved. Is there anyone out there monitoring this?
  • Which includes: Does anyone track the continued alignment of the objectives to moving targets?
  • Which include-leadsto: Does anyone track the continued alignment of all organisation incl RM activities to the changed objectives?
  • Do all truly recognise, in every their daily activities that actually they contribute towards some defined objective(s) ..? Don’t lie;
  • Do all that deal with ‘controls’ recognise, in all their work [i.e., including that their very presence is a control of a slightly higher level only] that they deal with Whats that are a translation of Hows of a higher level, and their Whats will be translated into lower-level (higher!-detail) Hows, and that it is their duty and obligation [failure of delivery upon which they are to be sacked with full salary clawback, obviously] to take into account where the ‘downward’ translation leads to, all the way to shop floor level, and see whether the controls there are still workable or else the translation needs to be re-done until this is solved? Don’t lie;
  • Do all involved in RM, see that they’re dealing with preventing negative influences from an ever-changing context, in the hope that the positives or even merely the neutrals, aren’t prevented by the very ‘controls’ they so desperately want [in lieu of understanding that one mustn’t want control but positives]? I repeat;
  • Returning to the above-mentioned piece: Do all involved in RM understand [which of course goes waaay beyond mere knowledge of the regurgitating kind that most risk managers I know are at max capable of] that each and every choice, in goals, objectives, and controls, means alternatives are [at least partially] stalled or foregone ..?
  • [Oh and don’t forget that when setting goals, and subsequent objectives, that the goals aren’t really Capital’s anymore, since decades or ever; they’re more and more clearly, society’s wishes as license to operate for you.]

I.e., we need real options back in the game, with a portfolio (management) approach – weighing rewards, and risks, in a balanced [and capped] non-perfect correlation manner that recognises that not only risk ceilings but also negative correlations play, here. Like, doing one thing foregoes 100% of the benefits of doing something else since you can’t do the latter out of budget constraints. Spend [‘invest’ with inclusion of your average success rate] on the one, cannot spend on the other. This also goes for spending on the one and not being able to afford sufficient buckets on risk control for the other [to manage that one back to acceptable residual risks].

There; you now see how interested I was in that other piece..? Tok about widening of your scope, being both peri-scope before launching your torpedoes, and scope being breadth, width of context and environment under study.

To ponder, I leave you with:

[ … del Sal certainly has seen its risks, all around and internally much less; Cataluña]

Bill the Kid no longer

Lots’a discussions again, lately, about accountants’ fees. Systemic problems, mostly. Fat rates, but when [i.e., not if] the CPA is found out to have done (much too) little and/or knew about the continuity in doubt, there’s the ubiquitous “I’m from Barçelona”.
Fat rates were there to cover the insurance to cover your inevitable (sic; risk-based decisions…) errors. Unless gross misconduct is in play, which shouldn’t be hard to prove. So you’re rewarded for what, actually ..?
Why press all the underlings so they resort to drugs (yes; not me I’ve never touched them, but know of scores who do), burn-out or similar excuses not to slave to death for your lazy (sic) benefit. Until you can replace them with ML – also then, you may need some re-adjustment … as e.g., here.
Being paid by the very ones you would call to account … Right; are you ignorant below kindergarten level or wilfully lying about your ability to remain independent and objective? … That reads a bit harsh, even more than I intended. To add, this (in Dutch): with domineering ‘governors’/executives, one needs of fact more strict enforcement. E.g., on the independence thing and external accountants being appointed not by / through (the near-binding advice of) executives but by the Board of Trustees that represent all stakeholders but not the executives.

Maybe there’s a road (avenue..?) to remain relevant, by instead of charging for any best efforts (which are almost completely none of the two), you’d start charging for (only part of the profit of) your results.
I.e., you earn 10% of e.g., savings your advice [which is most of what you’ll be left with … based on experience oh a loophole] delivers to the client. Needs a. an innovative accounting mechanism at your client, b. a client that happily forks out millions if you save hundreds of millions even if your advice only took you a couple of hours. a., I’m unsure about. b., I know from 2nd-hand experience (18 years back) is a psychological Problem, on the client side.

Oh well. We’ll keep on discussin’
Plus:

[A small overview of all the alarm bells and whistles that you should’ve gotten repaired at your client…; Noto, Sicily]

Gobbling up your capac – or not and do something useful

A somewhat cryptic title again.
This, about the latest strand of news (‘thread’ or worse..?) wherein ML is accused of gobbling up so much compute capacity that … Well, at least that the world will come to an end through global warming that was so much due to power consumption for ML. Oh no, that was ‘blockchain’ as the culprit, until recently. Where of course the solution was already to dumb down ‘chain-algo’s to a level that no-one would be sure that the self-righting ship [hey look that has its own wiki: here with nice pics to boat boot] mechanisms are still in operation. If they ever would have worked, in full public chains only, … if.

Now then, the ML part. As here. Given a. it’s historic data not logic, b. it’s nicely exponential, c. we all know that exponential is only a local characteristic ..! [yes, sic], d. meltdown hasn’t been upon us yet … we should conclude e. apparently, what ML finds in data has not necessarily any bearing on reality, f. we should do something; e.g., adapt our methods [which was over-, very over-due anyway ..!]:
Don’t rely on induction.
At best, rely on deduction [e.g., from business process analysis downwards ..!!] and possibly verify / falsify with data analytics. Analog to [but at a sharp angle with] this.
And/or look for hidden patterns in the data, but assume not that there are any, worthwhile ones ..!!
And if the latter, augment with pushing through induction that until here wasn’t anything of the kind, with analysis into theory/logic which would give you back the rules you’d need to turn the whole thing into either expert systems [the fruitful path] or episodic learning [the haphazard amateurs’ path].
And, at the very least, do not I repeat do not do any analysis before you have clearly established you have the right data for your problem. If you don’t know your problem, woe you. Turn to At best above.
And, realise that the future production runs will be on – relatively …! – moderate flows of transactions through your system, not the ginormous bulk of bwarp that you train the system on.
And remember this.
Then, your capacity problems will not aggregate to a global problem.

Also, isn’t this quite the similarsame as what happened say, a decade ago in Storage ..? Demand rose out of control, and couldn’t be supplied with sufficient supply [hmm, semantic ‘logic’] to even keep up to keep the gap (linearly) as it was.
Nowadays, is that problem still here ..? What happen ..?

So yes, ML causes 99½ problems. Compute capacity is ½ one.

For now:

[Ask yourself: Do you feel 5D-computy today, punk ..!? Do You …!?; Toronto]

No fools – but paid like ones

Before you’re off to celebrate the 11th of the 11th – not as much the /un/official start of carnaval in the Dutch bayou but St. Maarten’s day, a sobering thought:
You thought that all research showing robots will not take away our work but change it or we ourselves evolve our jobs to something unrobotisable and gloriously creative and art’sy, were and are true.
You’re not mistaken.

But …

Also, you are. The future isn’t that bright. Though you may have a job and all, it’ll not be and all. As shown in this scientific paper, you will be degraded. And paid less; the dent will remain as a scar on your working life balance…

There. Or, you prepare by screamingly fast evolving your work capabilities to deal with such a future.
On that happy note, I can leave you with:

[You wanted the fog to clear. To a clouded sky. Well done; Villa Lusthusporten Stockholm but you disconnected at the first 4 letters of that name similar to in the above story]

Titulaer was ahead of most of us …

As @Schellevis tweeted, this old find in a (pre-)junior high magazine of, mind you, 1982, the then often ridiculed Dutch version of Carl Sagan / Neil deGrasse Tyson with the original below, and I quotranslate: <
* What will our life look like within ten years ..?
I think it will hardly take ten years before every household has a computer. Remember that in most households, there already is one. Those little computers are of course very small, built in into all sorts of devices. There’s a lot of washing machines with built-in miniature computers. There’s intelligent TVs with which you can do allsorts. There will probably be some form of computer that will take over many tasks. That will control the central heating, that will check whether all doors are locked at night; that sort of thing. That’s one aspect of the future. A second aspect that is very clear, is that screens will turn up everywhere.
* Oh really?
In much more places than today. They’ll not be TVs only anymore. But you’ll see who’s at your door. You’ll be able to see at what time the bus will be at the stop, and what discounts your grocer will have. Also, the screens will be both much larger and much smaller. There will be wall-to-wall screens. On such screens, one could play video discs, etcetera. I also think a third thing that will clearly change in the future is that gadgets will be much more user-friendly. I think you’ll be able to talk to chips, and the chips will listen and talk back.
* How’s that going to work?
Like this: The TV remote will disappear. If you want to switch to the second TV channel, you’ll just say “channel two”, the device listens and you’ll have channel two. If you say “brightness” it will turn up brightness. You’ll have this in your car, too. At some point there will come a device that tells you when you have to pull over for gas. If you don’t, it’ll tell you “doofus, you’ll run dry soon”. There are watches today that wake you up in the morning. I have a clock at home that tells me to get out of bed.
* If we all have Cable, much more is possible isn’t it?
One could build a cable network over which gas and electricity utilization is reported. But also you can order groceries over it. I think there will be somewhat more home delivery. Linked to this is of course payments, because banking runs over the same system.
* Oh but our money will disappear then, won’t it?
Yes, money will almost completely disappear. And with that disappearance, there’s a fun thing to mention. Being these electronic payment cards. I have a mock-up of one at home. In those cards, a mini-computer is built in. With such a card, you could e.g., pay for the payphone, or for the bus ride, or for parking. You might even plug the card into your TV to check your bank balance.
* Droves of satellites are sent into space. What will they be used for?
Already, they enable us to watch live TV. But they can go way beyond that. With a satellite one could develop a system by which you can track where everyone is at all times. You can see where they are in their car, but also at home. One could make phones that can be built into wrist watches, and then you can be in contact with anyone around the globe via satellite connections. From the satellites you could see whether there would be an oil spill anywhere, and track and arrest the perpetrator within five minutes.
* That sounds rather threatening, devices that can talk to you, everyone can see you …?
Yes, but you’ll have to learn to live with these things. I’m not overly happy with all developments myself. But I do keep an eye on them.
> unquote.

Yes, prof(?) Titulaer PhD was scoffed at. Tiny-partly because a number of his predictions didn’t pan out in the [shorter of the] time frames he set, but mostly because of his looks. When reading the above, the word ‘vindication’ is the least we can apply.

Now then,

[Uncanny dystopia already back then / now …; Zuid-As Amsterdam]

Oh and the original:

Software – Museum Stuff OR Events

In a time long forgotten (i.e., two years ago), on a ‘net far far away (i.e., hopcount > 2), I had this post about how, besides the hardware museums that exist here and there [rather sparsely, given the prominence of the stuff], there should be software museums as well, of some sort [on the ‘net itself? may not stand the test of time; also think of mere power failures and insufficient back-ups, or political history re-writing – the ‘klout’ is only someone else’s computer…].
Plus of course some museum for algorithms, due to this – algos can be well thought out true pure art, or constructed messily. The art ones stand out, but still isn’t it an art to make them; including possible sub-extreme-efficiency-brushing that compiler optimisers do, being dull and not quite the perfect off-perfection [perfection is Boring, remember?] that real Art is.

Now, I stumblebrowsed [pleonasm alert] upon this Slate article, on straight lines of code not even the emerging property of ‘algorithm’ that in a strange twist existed before the medium… I mean; one implements a nous of the function into lines of code that, when run, demonstrate the function i.e., are the algorithm but the algorithm and the function and the nous are different things aren’t they? … hard to explain this mix-up, seems to be outside the possibly transcendental phenomenological apparatus or is it?

Yes. We should see such a thing. Also, the above. Plus:

[Yes my wife [that’s actually her] knows Art; The Hague – what’s with the ridiculous museum name change ..!? Phone pic so don’t bother commenting on the quality]

Some balance in wine appreciation (Dutch)

Even de balans zoeken, in wijn proeven.
Eerder had ik al deze met een incompleet overzichtje van proef’methoden’/scores. Waarbij de basic chemistry, en het vastgestelde feit dat de tong hoewel per individu sterk variërend in ‘specialiteit’- en gevoeligheidsprofiel (4D) best wel even of veel meer gevoelig is dan ballpark lab-meetapparatuur, dus absoluut waarde heeft om aan te tonen dat iets met graniet ook daadwerkelijk die sensatie (sic) geeft en een hint van blauwe bessen dus betekent dat er een ietsje blauwe-bessenchemicalieën in zitten!

En met een uitstapje naar kaas. Zowat te vanzelfsprekend om nog te hoeven noemen.
Waarop bovendien een sausje van Beilagen in het algemeen in dit stuk. Oftewel; een diner moet [voor mij maar hé ik schrijf dit dus ik mag schrijven wat ik wil] 50-50 eten-wijn zijn, qua kwaliteit van de gang op bord en in het glas. Niet qua volume; kleine glaasjes want proeven boven drinken! en dat betekent dus ook dat de kwaliteit per volume in de wijn veel hoger moet liggen dan op het bord.

Nu is er dan eindelijk weer eens pertinente nieuwe info, wetenschappelijk gevalideerd dus: de op het oog belachelijke beschrijvingen van sommige kenners [als zodanig breed erkend…], worden door andere kenners juist geïnterpreteerd. Het is dus een kwestie van non-, semi- en vol-verbale signalling voor de in-crowd. Waarschijnlijk niet bewust, wellicht zelfs als jargongebruik-drempelverlagende inzet. Want als het volslagen verzonnen klinkt, kan iedereen er dus een eigen draai aan geven en niks is fout. Hetgeen allemaal heel juist is. Als er maar niet wordt geroepen ‘dat mag je niet proeven want dat staat niet in mijn WSET3-lijstje dat ik uit het hoofd heb geleerd ook al proef ik er niks van’.

Nou ja. Laten we het dus breed houden!
Tevens:

[Komt er inmiddels ook wijn uit de omgeving van Lake Geneva ..? Possibly.]