Disqus - Latest Comments for vidarh

Re: No Ghost in the Machine

Vidar Hokstad — Mon, 09 Mar 2020 11:24:48 -0000

Free will is the ability to choose between
alternatives, a process you consciously engage in all the time.

The very simplest computer can choose between alternatives, so clearly that is not the essence of any idea of free will.

That your side-excursion through 'randomness and deterministic processes' uncovers nothing but is hardly surprising.

You don't engage with it at all, you sidestep it, and avoid even trying to address the problem of how a decision-making process can exist without being a combination of deterministic and non-deterministic processes *or* explaining how either deterministic (that is, purely mechanical cause-effect) or non-deterministic effects (that is: those that can not be predicted based on the preceding events) can provide any kind of agency.

For whom or what could free will be an 'illusion,' in the absence of an experiencing subject free to meditate on the difference?

Why do you think a subject requires free will to meditate on anything? There is no reason for that to be the case.

In a universe bereft of emergence, including feedback loops enabling processes to self-adjust, how could the concept of 'agency' even arise?

This question does not make any sense. 1) I have not suggested a lack of emergence or feedback loop - both exists perfectly well in both systems with purely deterministic behaviour (e.g. Cellular Automata and flocking behaviours are perfect examples of emergent behaviours and feedback loops in such systems that are well studied in computer science).

2) My point is exactly that agency makes no sense. You're asking me to argue against my own point here.

How do you account for this knowledge of yours, and for the manifest differences between arrangements of matter that give rise to consciousness, and arrangements of the same matter that do not?

Why do you think agency and free will has any relevance whatsoever on this?

Explain also, if you can--without resort to question-begging--just how the existence of agency in the cosmos offends logic.

First define agency for me in a way that does not simply devolve to obscuring determinism and probabilistic randomness behind a layer of waffle. How is it separate?

It's not my job to disprove something you've not even been able to come up with a definition of. Define it first, then we can discuss whether it can exist.

My entire argument is that you will not be able to come up with a definition that can not be stripped bare as nothing more than a combination of determinism and probabilistic randomness, or shown to be impossible.

Your reply has me wondering how often in your life you've opened books of logic... do you even know the difference between Modus Ponens and Modus Tollens? What incompatibility, either conjunctive or disjunctive, do you imagine exists between agency and logic, and what connection does it have with magic, or 'black boxes?'

This is cute. I challenged people for a definition, and you fail to come up with one and asks me to disprove a definition you have been unable to provide.

[Edited to add:] Do you honestly believe people aren't responsible for their actions?

The flaw in this is that "responsible" is a vague term with many different meanings.

Do I believe people are *legally* responsible? Yes, I do, because legal responsibility has nothing to do with whether or not the persons involved has any actual influence on a molecular level over the decision making process that leads to the specific actions.

Do I believe people are "responsible" in the sense that they have agency and that we add something to a decision that is not a pure cause and effect or a probabilistic randomness (e.g. at quantum level)? No, I don't, because nobody has ever been able to even define what that would be. You didn't even try, just resorted to handwaving and trying to get me to define it for you by trying to get me to argue against a concept you've not defined.

The challenge here for most, and I sense for you too, is that people tend to assume that if we stop believing there is agency, we'd just stop acting, but that makes no sense.

If I stop eating for example, I will starve. It doesn't matter that when I eat, my "choice" to do so does not involve agency in any meaningful way. It feels like a choice, and if I make the choice to eat I will be full afterwards. Whether or not that choice is an exercise in free will or agency does not change the outcome. And so I eat. And go to work. And in ever aspect act as if I had a choice.

The test of free will comes down to the belief of this: If you could roll the state of the world back to a point where you "made a decision", and everything, down to quantum mechanical effects and random outcomes, was the same, do you believe you could choose something else? (because if we do not assert the same physical state and same random outcomes, then all you would be demonstrating would be cause and effect and random effects)

Free will and agency would need to exist in that space: A "something" that would allow you to bypass deterministic cause and effect and bypass randomness, and somehow make a different decision without affecting any neurons in your brain or anything else mediated by physical effects (because otherwise it'd change the preconditions), and change the following brain state in ways that violate the laws of physics.

If you can define that "something" in a way that makes any kind of sense, you'll have the basis for a revolutionary paper in philosophy *and* the basis for Nobel in physics as such a violation of cause and effect would necessarily alter the state of matter in violation of the laws of thermodynamics.

Re: No Ghost in the Machine

Vidar Hokstad — Mon, 09 Mar 2020 10:53:24 -0000

The point is that all of this boils down to either devolving into arguments about computability, or defining it in a manner that circumvents the issue of computability and in doing so implicitly concedes that there is no way out that doesn't devolve to some combination of determinism and randomness that leaves no room for *agency*.

The very concept of free will in terms of actually talking about how a decision is made is meaningless, quite literally.

Re: No Ghost in the Machine

Vidar Hokstad — Wed, 04 Mar 2020 10:07:43 -0000

Try to define free will; I dare you.

The point made above is that there is no definition of "free will" that does not eventually devolve to some mix of randomness and deterministic processes, merely sufficiently obscured to be able to declare that at its core there is some magical "black box" imbuing some form of agency somehow outside logic that we're not allowed to peer into - because if we were, we'd again find some mix of randomness and deterministic processes.

"Free will" is simply a way of obscuring a process to complex for us to understand the mechanism, and an illusion we hold onto because the consequences of giving it up would be too troublesome (the morality of judging people for actions they have no agency over, for example, becomes dubious at best).

Re: Ruby, I love you but you're bringing me down

Vidar Hokstad — Mon, 12 Aug 2019 07:13:55 -0000

Consider "(", "," and ")" as operators too, and your last examples makes sense (though I agree with you the lack of a spec is painful - the de-facto precedence of "," and parentheses relative to the other operators isn't even mentioned in the Ruby core precedence documentation). The priorities of the operators are such that the closest valid parse for this:

some_method(a or b)

is still this:

some_method(a) or b

So "or" is only ever valid inside a parameter list if wrapped in another expression.

(and when you introduce a space, the parentheses implicitly become part of the expression for the argument rather than wrapping the argument list, which is why `some_method (a or b)` works).

Re: http://www.hokstad.com/mini-reviews-of-19-ruby-template-engines.html

Vidar Hokstad — Sun, 03 Sep 2017 21:28:51 -0000

As many years old as this article is, my comments on HAML above pretty much sum up my reaction to Slim: Significant whitespace makes me discount any language right away, I'm afraid.

Re: Inversions (Iain M. Banks)

Vidar Hokstad — Mon, 29 May 2017 11:03:33 -0000

Hi Leila,

I think the question is what aspect of it you liked, as it is quite different from the other books by Iain M Banks. The other books in the series are much more obvious science fiction. If you like science fiction I would suggest Consider Phlebas or Use of Weapons. Both will give you new insight into Inversions, as the two outsiders in Inversions are representatives of Contact and Special Circumstances respectively - two organizations in the Culture with different philosophies on how to deal with "more primitive civilisations". Inversions was on purpose meant to fit into that universe but not explicitly feature it.

If you like the adventure/fantasy aspect more, then perhaps Against a Dark Background or The Algebraist would serve to bridge the gap more. Neither are part of the Culture series - they are both standalone novels. Against a Dark Background is a fantasy / SF hybrid of sorts (there's technology, but it's not driving the story for the most part) set on a planet that is isolated in a region of space with no other star system within reach. So they're a technological society where a space race makes no sense. The Algebraist on the other hand is set in a star system that used to be part of a vast interstellar empire, but has been left isolated for two centuries after the destruction of its gateway to the nearest planet - two centuries being the time it takes to have a replacement sent to them. It has plenty of aliens and technology, but focused more on the adventure of the main character while trying to get to the bottom of a mystery around the time the replacement gateway is meant to arrive.

Matter might be another option. Matter is a Culture novel. It's later in the series, but chronology does not matter much for the Culture novels. It's set on an alien world where a pre-industrial feudal society is about to get into contact with the surrounding world. It's a related theme to Inversions in a way, but much more explicit (the Culture parts of the story are explicitly told). Matter is not my favourite Culture novel, but it's still worth a read.

Re: http://www.hokstad.com/5-simple-ways-to-troubleshoot-using-strace.html

Vidar Hokstad — Wed, 19 Oct 2016 11:07:53 -0000

I don't think he's copied - there are enough differences there that it could just be coincidences. But thanks for mentioning it anyways - I'm amazed this page still draw in people 8 years after it was written..

Re: Upcoming Marvel Movies: Phase 3 Title List And Release Dates

Vidar Hokstad — Thu, 08 Sep 2016 13:53:06 -0000

Heh. "Am ..." instead of "I am" is a Nigerian thing, and +234 is the country code for Nigeria. It's a whole new fun variation of the 419 scams... Yay, just what we need.

Re: Writing systemd Units 2016-06-23 By Vidar Hokstad

Vidar Hokstad — Thu, 28 Jul 2016 12:36:10 -0000

Thanks - very interesting article. Socket activation in particular isn't covered much elsewhere, and while I tend to use Docker I love seeing more coverage of systemd-nspawn.

Re: Writing systemd Units 2016-06-23 By Vidar Hokstad

Vidar Hokstad — Thu, 28 Jul 2016 12:30:56 -0000

You're welcome. Glad it was helpful

Re: GitLab Container Registry

Vidar Hokstad — Fri, 03 Jun 2016 10:31:43 -0000

How's that going? I'm debating whether to actually set up local runners or wait for your shared runners to handle this (love the container registry - that finally got me to register; not quite ready to ditch Github for my public stuff, but suspect I'll be migrating my private stuff off Bitbucket soon...)

Re: HC | Welcome

Vidar Hokstad — Wed, 27 Apr 2016 06:51:54 -0000

http://hokstadconsulting.co...

But thanks for reminding me to add an auto-discovery header.

Re: Aldri mer Tsjernobyl

Vidar Hokstad — Mon, 25 Apr 2016 09:19:05 -0000

De kunne godt tatt med vannkraft og solcellepaneler på den oversikten også - det ser heller ikke bra ut i forhold (dammer kollapser; folk dør under installasjon av solcellepaneler - konstruksjonsulykker utgjør en stor andel av dødstallene for kraftverk, og solceller trenger langt mer konstruksjon per enhet energi).

Vind er muligens tryggere, og det er mulig at enkelte former for sentraliserte solenergi-installasjoner også er tryggere - jeg vet ikke om det er nok tall enda til å gi klarhet i det. Men totalt sett er det svært klart at atomenergi er et av de tryggeste alternativene vi har, og kan bli langt tryggere (dødstallene vi har er basert på tiår med erfaring med reaktortyper som alle er utdaterte)

Re: Aldri mer Tsjernobyl

Vidar Hokstad — Mon, 25 Apr 2016 09:12:09 -0000

Og det sier i så tilfelle sitt om hvor useriøse de fleste norske miljøorganisasjoner er, i og med at av alle formene for kraftverk vi har i dag så er atomkraftverk blant de aller tryggeste, og har forårsakt langt færre dødsfall en de fleste alternativene per enhet energi, og er også ganske så langt nede på listen over negativ innvirkning på miljøet.

Re: Aldri mer Tsjernobyl

Vidar Hokstad — Mon, 25 Apr 2016 09:03:32 -0000

Poenget er at du kan ikke stenge disse kraftverkene uten et alternativ. Dersom alternativene er noe annet enn et nytt atomkraftverk, så er sjansen stor for at det øker oddsene for dødsfall betraktelig.

F.eks. pumper kullkraftverk ut massive mengder med radioaktivt materiale - mye mer enn de totale utslippene som følge av ulykker på atomkraftverk.

Vannkraft har drept mange ganger så mange pga. dammer som kollapser. Banqiao ulykken alene tok livet av 171.000 mennesker.

Flere dør som følge av ulykker under solcellepanelinstallasjon per enhet energi tatt ut enn det har dødd i atomkraftulykkene samlet.

Det er greit at enkelte reaktortyper bør erstattes. Vi har reaktordesign i dag som er utrolig mye sikrere (helt opp til reaktorer der reaksjonen bare stopper dersom systemene rundt feiler). Men så langt er atomkraft en av de aller tryggeste energiformene vi har. Selv Tsjernobyl hendte kun fordi operatøren utførte noen vanvittig uansvarlige eksperimenter.

Dersom man stopper disse reaktorene uten en klar plan for et bedre alternativ, så vil man ha liv på samvittigheten.

Re: Aldri mer Tsjernobyl

Vidar Hokstad — Mon, 25 Apr 2016 08:56:38 -0000

Du kunne spredt alt avfallet i atmosfæren, og atomkraft ville _fortsatt_ være en av de tryggeste energikildene vi har.

Problemet er lett å løse: Mer atomkraft. Det er reaktortyper som kan bruke mye av avfallet som drivstoff, og etterlate langt mindre farlig avfall, og nyere reaktortyper etterlater avfall som er langt enklere å håndtere.

Problemet med avfallslagring for atomkraft er først og fremst skremselspropaganda og politikk. Det materialet som er spesielt farlig utgjør en ørliten mengde masse som ville gjøre svært liten skade selv om det kom på avveie.

Re: Fascisme i USA? Ikke helt utenkelig

Vidar Hokstad — Wed, 16 Mar 2016 12:44:26 -0000

I etterkant ser i hvert fall ikke anklagene om at han truet verdensfreden så urimelige ut, med tanke på at den aggresive håndteringen av forholdet til Sovjet tidlig i presidentskapet hans nesten started atomkrig ved å skremme Sovjetiske ledere til å tro at USA planla å starte krig, og fikke dem til å seriøst forberede seg på å komme USA i forkjøpet:

https://en.wikipedia.org/wi...

Det var først etter at det han fikk vite om den Sovjetiske responsen på Able Archer i ettertid skremte vannet av ham at Reagen tok en mer forsonlig tone mot Sovjet.

Re: http://www.hokstad.com/rails-is-total-overkill-and-why-i-love-rack.html

Vidar Hokstad — Fri, 11 Mar 2016 20:43:29 -0000

Yeah, that happened long after this was written. As I noted in my reply to your other comment, it doesn't really address my issue with Rails, though. Rails of today *is* much better in almost every respect, but it is still for most practical purposes a monolith that makes a lot of choices for you in ways that makes it hard to pick and choose, even if it's a monolith packaged in pieces. E.g. trying to use Rails with a different ORM than ActiveRecord is "fun", because even if you get it working for the basics, if you want to actually take advantage of other code written for Rails, most of it tends to assume Rails with all default choices.

As I noted elsewhere, though, if anything my view is even more minimalist now:

This blog is now static pages (admittedly part of the reason that works is because I've outsourced my comments to Disqus). The code that generates the pages is much the same as it was. The views didn't change when I switched over. And I have a tiny little Sinatra app to let me view in-progress drafts on the fly. But the finished site is generated statically.

Today I see dynamically generated pages as a code smell except where there is a very clear need (e.g. rapidly changing information), and even then I'd consider edge side include options to dynamically generate the bare minimum. The cost of dynamic page generation outweighs the cost of static page generation in far more instances than people think. A *lot* of pages are reasonable to trigger generation of when data changes, or simply regularly rebuild, rather than dynamically generating pages on access.

This also tends to make heavy options like Rails even more unnecessary: 90%+ of the web apps I work on (and that includes a number of Rails apps) only need dynamic requests to handle API calls, where most of the type of functionality that Rails provides ends up going unused. A big benefit is content distribution and caching becomes vastly easier to do efficiently (e.g. it can make the difference between caching with the need to revalidate and statically serving up most or all content from the edge; something that makes a huge difference when you want to serve up content from servers on multiple continents, for example).

Re: http://www.hokstad.com/urls-do-not-belong-in-the-views.html

Vidar Hokstad — Fri, 11 Mar 2016 20:28:49 -0000

Looking at large web code bases regularly, I can tell you that even today, putting hardcoded URLs into views is more common than not, so I think this is about as relevant today as it was back then... The gap between common practice and best practice if anything has widened.

As for the Rails approach, I haven't kept up with how route helpers have evolved in the newest versions of Rails, but for years after I wrote that article at least, Rails was still subject to the same issues I elaborated on in one of my older comments, which meant most Rails apps at least then had massive dependencies between views and the routing configuration. What I advocated in the post was to take an approach suggested by Terence Parr, where the controllers would explicitly be left in control over what information is accessible to the views in a way that lets the views request "route for X" but where the renderer provided determines what "route for X" means. Doing so effectively decouples the routing and views. Asking for a link to "tag foobar" can mean something entirely different in the context of a users profile and the general site, for example, but the view does not need to know the difference.

The comparison between Rails and Rack now or calling Rails overkill may seem odd (it was controversial enough then), but remember that at the time Rails did not interface to web servers via Rack, and Rails was far more monolithic. Especially with the "Merb merge", Rails was decomposed far more. Rails of today is a lot more nimble than it used to be. I still consider it massive overkill, though. The point of the comparison to Rack was not to suggest that Rack was enough, but to point out a massive difference in philosophy: Rack explicitly aimed to decouple and make it easy to replace components, and make it easy to interject functionality (middleware). Rails, on the other hand, at the time (but to some extent still) aimed to include the kitchen sink and provide it all in one package. Rails of today still comes with the kitchen sink, but replacing parts of it is a bit easier, and it uses Rack to provide some degree of decoupling some types of functionality.

(incidentally, my blog is much lighter weight than it was 8 years ago: I ditched my web app for a static code generator - not only is Rails overkill, but increasingly I have come to see dynamically generated pages as frequently being overkill and an absolute last resort, and statically generate whatever I can get away with...)

Re: Why coupling is always bad / Cohesion vs. coupling

Vidar Hokstad — Fri, 11 Mar 2016 12:26:51 -0000

No, all classes don't create tight coupling in the sense meant here. Coupling is created by combining code that carries out different responsibilities in ways that makes it hard to separate them. If you write classes that have one responsibility, and you connect that code to other code in ways that do not hardcode a responsibility, then you don't create tight coupling in the sense discussed here.

The point is to avoid unnecessary dependencies, and to avoid hardcoding the dependencies you do create. E.g. don't add your own logging implementation to the class, delegate it to another class, and allow providing another log object. The former (adding your own logging code to a class whose primary purpose isn't logging, or hardcoding the log object to use) would be tight coupling. The latter is loose coupling. The amount of code is pretty close, and providing defaults can be fine, but the latter is e.g. far easier to test or reuse in other contexts.

Re: Carl I. Hagen vil åpne oljefondet for flyktningene

Vidar Hokstad — Tue, 01 Mar 2016 05:03:39 -0000

Nå er det mange partier som opp gjennom årene har ønsket å øke norsk bistand nettopp fordi det er billigere å forebygge enn å håndtere kriser. I alle år har nettopp FrP vært en av hindringene mot dette. Kanskje Hagen burde ha tenkt over konsekvensene av den bakstreverske politikken deres før.

Forslagene hans om asylmottak i land som Libanon vitner også om kunskapsløshet. I Libanon er allerede en av fire som bor i landet flyktninger. De har tatt imot flere i forhold til folketall en nesten noe annet land i verden. Ønsker han at vesten skal destabilisere flere land?

Re: Frp's soldater

Vidar Hokstad — Tue, 23 Feb 2016 13:58:40 -0000

Det har ingenting med norsk kultur å gjøre, men med symboler som har en tradisjon for å bli misbrukt av høyreekstreme grupper, og at Norge har en historie der slike grupper har stått for det meste av politisk motivert vold de siste 3-4 generasjoner.

Om "Odins soldater" ønsker å gjøre norske gater til et mer vennlig sted hadde de vært bedre tjent med å velge navn og påkledning som ikke drar paralleler til grupper som stod for 70- og 80-talls nynazistisk terror, inkludert ting som brannbomber mot 1. mai tog, drap og diverse voldelige angrep.

Kanskje de bare er en gjeng historieløse kløner som ikke har tenkt over det, men siden mye av rekrutteringsgrunnlaget er ytre høyre er det mer nærliggende å anta at det de ønsker er å skremme bort personer de ikke liker.

Re: Frp's soldater

Vidar Hokstad — Tue, 23 Feb 2016 13:48:24 -0000

Og for andre vil det være motsatt hvem som føles tillitsvekkende og hvem som skaper frykt. Det er nettopp det som er poenget med å dra fram at borgervern er et problem: De er ikke nøytrale.

Hvis de ønsket å være nøytrale hadde de ikke valgt symbolbruk som har en lang historie i Norden hos til tider voldelige miljøer på ytterste høyre.

Kanskje de har de beste motiver, og kanskje de vil oppføre seg skikkelig, men inntil det er demonstrert så er det mange som har god grunn til å bekymre seg for at ikke alle som blir med kommer til å holde seg i skinnet.

Re: Fjordman svarer på kritikken

Vidar Hokstad — Tue, 02 Feb 2016 11:52:00 -0000

Hvilke problemer er det vi ikke hadde? Terror? Muslimske terror-grupper har enda langt igjen før de tar igjen den terroren våre hjemmedyrkede ikke-muslimske terrorister har gjennomført i Europa i etterkrigstiden. Også i Norge.

Re: Cameron i knipe i EU

Vidar Hokstad — Fri, 18 Dec 2015 06:43:51 -0000

Jeg har bodd i England i 15 år, og erfaringen min stemmer ikke overens med det du sier i det hele tatt.

Ikke engang det aktivt innvandringsfiendtlige Migration Watch tør å bruke så overdrevne tall. Deres tall sier 800,000 polakker. Og ytterste høyre i Storbritannia er de eneste som tror at dette truer sosialsystemet her - polakkene som kommer til Storbritannia gjør det for å jobbe. Faktisk har netto-migrasjonen fra Polen bremset opp fordi mange polakker velger å reise hjem fordi lønninger i f.eks. bygningsbransjen i Polen har gjort det mindre attraktivt å bli i Storbritannia.

Velferdsordningene i Storbritannia er også begrensede nok i forhold til e.g. Skandinavia at om det var for å utnytte velferdsordinger de kom hadde det vært langt smartere å dra til andre land. Some Norge.

Den største trusselen mot velferdssystemet i England er det Konservative partiet, som gjør det de kan for å bygge det ned.

Britene så på innvandring fra Polen som en trussel helt til de fikk oppleve den. Så ble de vant til den, og så skiftet skremmebilde fra Polen til Romania og Bulgaria, mens de fleste her synes det er helt greit at late Britiske bygningsarbeidere har måttet skjerpe seg for å konkurrere med langt mer profesjonelle polakker.

Når begrensningen på for Romania og Bulgaria ble opphevet så campet journalister på Heathrow for å være de første til å ta imot de forventede enorme hordene. Som uteble. Rumenere utgjør nå ca. 160,000 og bulgarer ca. 60,000; igjen i følge Micration Watch.

Realiteten er at tullballet til Cameron er spill for galleriet. Cameron vet utmerket godt at innvandring fra EU har ført til ytterst minimale probllemer.

Men det betyr også at han kan forhandle fram noe som høres ut som en seier. E.g. de ande EU landene bryr seg om fri flyt, ikke detaljene. En løsning kan f.eks. være at hjemlandet fortsetter å betale for velstandsytelser de første månedene. Det vil ha liten praktisk betydning, men ville gi Cameron nye argumenter for å bli.