[continued from previous message]   
      
   >> regular way, or to put it simply science assumes that "evidence means   
   >> something", i.e. empirical observation is an epistemologically valid   
   >> source of knowledge.   
   >> Occasionalism is the position that there are no natural processes;   
   >> instead God does everything. Progressive creationism can grade into   
   >> occasionalist evolutionism, but God, episodically or steadily,   
   >> magicking new species into existence is not occasionalist.   
   >>   
   >> Philosophical naturalism is the position that nothing is supernatural;   
   >> occasionalism is the position that everything is supernatural. Most   
   >> religious views lie somewhere between those two extremes.   
   >   
   > Most Christian views require at times irregularity (occasional   
   > occasionalism?), however they not antithetical to the pursuit of   
   > science. As I've said here before, identifying and demonstrating that   
   > boundary would be where science itself points to a creator. bmmmm   
      
   I believe that the usual phrasing would be "an interventionist god".   
      
   Did you intend to claim that the stochastic nature of radioactive decay   
   is points to a creator?   
      
   Your first problem is distinguishing unrecognised regularity from   
   irregularity. Your second problem is in leaping from irregularity to   
   creator. An unknown is not necessarily supernatural, and the   
   supernatural is not necessarily a creator god. Your third problem is   
   that this looks like an appeal to "the God of the gaps".   
   >   
   >>>   
   >>> The nature and measurement of information seems slippery. As you   
   >>> mention, is it Kolmogorov complexity or Darwkin's "incomplete record   
   >>> of the historical environment", or something else?   
   >>>   
   >>> ID posits a lawlike conservation of information, which I find   
   >>> intuitively appealing, but Dembski's efforts to formally define this   
   >>> have yet to land it seems.   
   >>>   
   >>   
   >> My intuition goes the other way. Hardware random number generators   
   >> create information out of "nothing".   
   >   
   > But not information that is specified (as in CSI), which is the critical   
   > distinction, and why Dembski and others give this so much attention.   
      
   That's a substantial retreat from conservation of information.   
      
   It seems to me that by claiming the genomes contain complex *specified*   
   information you're assuming what you're trying to prove.   
      
   >   
   >>   
   >> If the laws of physics are invariant with respect to time-reversal   
   >> then information (in some senses) is conserved. But while T-violation   
   >> has not been observed, physicists believe that the laws of physics are   
   >> CPT- invariant, and as CP-violation has been observed this implies   
   >> that T- violation also occurs. There is also the black-hole   
   >> information paradox, wherein black holes appear not to conserve   
   >> information.   
   >>   
   >> Creationists have been known to argue that evolution is impossible   
   >> because of the 2nd Law of Thermodynamics, ignoring that the law does   
   >> not preclude local decreases in entropy. (If the creationist 2LOT was   
   >> true, life would also be impossible.) Similarly, even if an analogous   
   >> law of information existed it would not preclude evolution (and life);   
   >> just as life (and evolution) exports entropy into the environment,   
   >> they could import information from the environment. Dembski et al   
   >> could retreat to the question of the ultimate source of the   
   >> information, but that is just the cosmological argument redux, and not   
   >> an argument against the factuality of evolution.   
   >>   
   >> For an analogy, consider the connectome - the set of connections   
   >> between neurons. In the same was as DNA this can been seen as   
   >> containing information. In most animals (C. elegans is an exception)   
   >> this is not fully defined by the genome. So a proportion of the   
   >> information in the connectome must be imported from the environment   
   >> (whether sensory inputs or biochemical noise).   
   >>   
   >> Turning again to the question of conservation of information.   
   >> AlphaZero, starting with nothing more than the rules, bootstrapped   
   >> itself to superhuman levels of play in, inter alia, Go and chess. Did   
   >> that process increase information? In that case where did the   
   >> information come from?   
   >   
   > As I’ve mentioned on t.o before, in the past I worked as an engineer   
   > programming Field programmable Gate Arrays (FPGAs). At the time, I read   
   > an article which utilised a genetic algorithm to develop an FPGA circuit   
   > for a clocked counter of some sort. It turned out to be a very efficient   
   > solution, but humanly incomprehensible because it appeared to utilise   
   > parasitic capacitances or some other secondary analogue effect. The   
   > device used was programmed with a 2kbit configuration file, and so it   
   > appeared that 2,000 bits of information (presumably qualifying as   
   > complex specific information) had been created de novo by an   
   > evolutionary process. I wrote to William Dembski at the time, who   
   > responded with an interest in investigating the example further, but   
   > offered an initial assessment that information had been “smuggled in” to   
   > the system by an intelligent designer (i.e. the creators of the   
   > experimental set up).   
   >   
   > My point being (without claiming anything definitive) the information   
   > may be from sources such as the intelligent system/algorithm designers,   
   > or from a brute force search of the entire soluition space (or enough of   
   > it to outperform humans).   
      
   The whole point of using genetic algorithms is that, in suitable   
   domains, they greatly outperform brute force searches. They are   
   vulnerable to hanging up on local maxima, but this can be in part   
   addressed by annealing.   
      
   I find the "smuggled information" response underwhelming. If a genetic   
   algorithm can import information from an "artificial" environment set up   
   by human experimenters then why can't it import information from a   
   "natural" environment? At which point you're back at the cosmological   
   argument again.   
   >   
   >>   
   >> (AlphaFold broke the back of the protein folding problem, but in that   
   >> case one could appeal to import from environment as the source of the   
   >> information in the trained model.)   
   >>   
   >   
      
   --   
   alias Ernest Major   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)