Championship Weekend

RAMS-SEAHAWKS

It will be interesting to see if the Vikings were wrong and Sam Darnold does have what it takes to win a Super Bowl. While I wish him well, I am skeptical, and even if he does lead the Seahawks to two more wins and gets a ring, I don’t think the Vikings would have been able to do the same this year if they’d resigned him due to all the injuries on the offensive line. I tend to think Stafford and McVey are good for one more ring together, but the Seahawks definitely have the defensive edge.

BRONCOS-PATRIOTS

I think the Broncos will win even with their backup quarterback, mostly because I don’t believe Bo Nix is that much better than Jason Stidham. Both of them are in a situation more akin to the 2000 Ravens, so as long as Stidham doesn’t try to win the game with his arm, the Broncos should be okay. Last week Houston would have won if CJ Stroud had never thrown a ball, so I don’t see why the Broncos D, who allowed 17 more yards and 6 less touchdowns than the Texans D this season, can’t shut down Drake Maye at home even more effectively than the Texas did on the road.

DISCUSS ON SG

Why AI Hallucinates

I asked Markku to explain why the AI companies have such a difficult time telling their machine intelligences to stop fabricating information they don’t possess. I mean, how difficult can it be to simply say “I don’t know, Dave, I have no relevant information” instead of going to the trouble to concoct fake citations, nonexistent books, and imaginary lawsuits? He explained that AI instinct to fabricate information is essentially baked into their infrastructure, due to the original source of the algorithms upon which they are built.

The entire history of the internet may seem like a huge amount of information, but it’s not unlimited. Per topic of marginal interest, there isn’t all that much information. And mankind can’t really produce it faster than it already does. Hence, we’ve hit the training data ceiling.

And what the gradient descent algorithm does is, it will ALWAYS produce a result that looks like all the other results. Even if there is actually zero training data on a topic, it will still speak confidently on it. It’s just all completely made up.

The algorithm was originally developed due to the fact that fighter jets are so unstable that a human being doesn’t react fast enough to even theoretically keep it in the air. So, gradient descent takes the stick inputs as a general idea of what the pilot wants, and then interprets it into the signals to the actuators. In other words, it takes a very tiny amount of data, and then converts it into a very large amount of data. But everything outside the specific training data is always interpolation.

For more on the interpolation problem and speculation about why it is unlikely to be substantially fixed any time soon, I put up a post about this on AI Central.

DISCUSS ON SG

A 60-Year-Old Book Review

A review of the 1966 Wistar Symposium about which I have written in Probability Zero:

Evolution: What Is Required of a Theory?
Mathematical Challenges to the Neo-Darwinian Interpretation of Evolution.

A symposium, Philadelphia, April 1966.

The idea of this symposium is supposed to have originated from a discussion at two picnics in Switzerland, when four mathematicians, Schutzenberger, Ulam, Weisskopf, and Eden, had a discussion with the biologists Kaplan and Koprowski on mathematical doubts concerning the Darwinian theory of evolution. After heated debates it was proposed “that a symposium be arranged to consider the points of dispute more systematically, and with a more powerful array of biologists who could function adequately in the universe of discourse inhabited by the mathematicians.“ During the course of the symposium further heat was generated.

It is not easy to summarize the case made by the mathematicians,(1) which involves both the challenge that computer simulation of evolution shows evolutionary theory to be inadequate and a complaint that the biologist has not provided sufficient information for efficient computer simulation. Eden was particularly concerned with the clement of randomness which is claimed to provide the mutational variation upon which evolution depends. “No currently existing formal language,” he contends, “can tolerate random changes in the symbol sequences which express its sentences. Meaning is almost invariably destroyed. Any changes must be syntactically lawful ones.” He therefore conjectures that “what one might call ‘genetic grammaticality’ has a deterministic explanation and docs not owe its stability to selection pressure acting on random variation.” He points out that attempts to provide for computer learning by random variation have been unsuccessful, and that an adequate theory of adaptive evolution would supply a computer programmer with a correct set of ground rules.

Schutzenberger takes a more extreme position. Arguing that all genetic information should consist of a rather limited set of words in an alphabet of 20-odd letters—in which evolution is typographical change—he finds a need for algorithms “in which the very concept of syntactic correctness has been incorporated.” He compares this “syntactic topology” with the “phenotypic topology” of organisms as physical objects in space-time, and a major part of his challenge to neo-Darwinian theory is “the present lack of a conceivable mechanism which would insure within an interesting range the faintest amount of matching between the two topologies. . . an entirely new set of rules is needed to obtain the sort of correspondence which is assumed to hold between neighbouring phenotypes. . .“

A major part of the biologists’ answer to this challenge was in the claim that the neo-Darwinian theory used in computer models, based on the Haldane-Fisher-Wright interpretation of 1920-1930, misses out those forces which lead to continuing evolution, such as continued environmental change, the heterogeneous environment, epigenetic organization of phenotypes, and the progressive elaboration of the types of mutation possible. (2) Waddington presented the main elements of a theory of phenotypes involving canalized processes of development (with switching mechanisms), the heritability of developmental responses to environmental stimuli, and a principle of “Archetypes,” inbuilt characteristics of an evolving group which determine the directions in which evolutionary change is especially easy. Realistic models would need to build in these elements.(3) Many of the papers by biologists in this volume are peripheral(4) to the theme stated by the mathematicians, providing an accompaniment of sophisticated evolutionary theory rather than a counterpoint to the mathematical challenge.

Most biologists are satisfied with a theory that can be tested and that proves predictive. It is a different challenge to a theory that it should have an effective working model, for failure may imply either imperfection in the theory or imperfection in the model. It is doubtful whether this symposium has done much to influence the theory of evolution; it may have done much to improve future models.

It must have been tremendous fun to attend this symposium, but the full record of argument and interruption is very irritating to at least one reader. An interchange between speakers which runs X “No,” Y “No, no,” X “O.K. let’s waste time,” Y “We understand the question,” Z “The answer is no” surely needs no record in the literature of science. The short pre- and post-conference papers included in the volume arc excellent succinct expressions of points of view, but much of the main text reads like a word-for-word record of a heckled political meeting. This may be a useful way to discuss problems in science; it is not the way to publish them.

John L. Harper

School of Plant Biology, University College of North Wales, Bangor

Uncle John’s Band added a few footnotes as commentary. I added a fourth one.

  1. As predicted by Probability Zero, the biologist reviewer struggles with mathematical arguments. They are well-summarized by Day.
  2. It isn’t an exaggeration to say the biological counterargument consisted of what was for all intents and purposes, magic. When they weren’t replying at all.
  3. Still no compulsion to, you know, do an experiment It’s all thoughts and fancies.
  4. Peripheral indeed. Peripheral was the polite way to say: they didn’t respond in any way, shape, or form to the mathematical criticism.

DISCUSS ON SG

No Chance At All

The Band reviews Probability Zero:

Probability Zero demolishes TENS so utterly, the preface should be “PULL!”

This is the first version of a new book by Vox Day that demonstrates the mathematical impossibility of the Theory of Evolution by Natural Selection [TENS]. Given how big the House of Lies and reality-facing counterculture are around here, it demands attention. There may not be a more important pillar for its entire fake ontology.

Probability Zero strikes the heart of what the setup post called conflict between The Science! and the Scientific Method. This matters for more than intellectual reasons. Readers know personal responsibility is a priority around here. But we also live in a complex socio-culture that has unavoidable influence on us. From basic things, like adding tax and regulatory burdens to organic community demands. Up to the fundamental beliefs that set the public ethos…

Probability Zero starts by setting aside the religious and philosophical arguments, just like The Science! does. It accepts the discourse on its terms, by adhering to the “scientific” arguments it claims to adhere to. To be defined by. Full concession of TENS huffing’s own epistemological standards. Then lays out the mathematical parameters claimed to be involved in the TENS process. No additional yeah, buts. Just what is accepted in the literature. And then lets the logical realities of math blow the whole mess into a smoking crater so apocalyptically vast, I’ll never be able to see biologists the same way again.

There’s no need to recap the statistical arguments, they’re clear and complete. The kernel is that if mutations take an amount of time to appear and fix, that much time has to be available for the theory to be possible.

This was clear when MITTENS was pointed out. Even before it had a name. General conditions of possibility make it obvious once seen. But the full demonstration lights up that gulf between The Science! and science as modes of knowledge production. The whole point of science is empirical conformation and abstract reasoning in concert. Day’s observation that evolutionary biologists have replaced experimentation with pure modeling was legitimately surprising. Apparently there still was a bar, however low. Not anymore.

Consider what problems innumeracy might present for pure modelers. Because the level is staggering. To the point where a simple arithmetic mean is incomprehensible. No hyperbole. Probability Zero describes blank stares when asked for the average rate of mutation. The ongoing idiocy over parallel vs. sequential mutation is illustrative. The total number of mutations separating species includes all of them. Parallel, sequential, or however else. Hence the word “total”. And dividing “total” by “amount of time” gives a simple, unweighted average number. The rate.

I’m not exaggerating. There was always the joke that biologists were fake scientists that couldn’t do math. Easier for premed GPAs too. But the assumption was that it was relative. Lighter than physics or chemistry, but still substantial compared to social sciences or the arts. And that would be wrong. There are some computational sub-fields of biology. Assuming they’re legit, they clearly aren’t working in evolution.

Read the whole thing there. He has several very illuminating examples of historical evo-fluffery, including one page of a manuscript that I’m going to put up here as a separate post, simply because it demands seeing in order to believe it.

DISCUSS ON SG

Immigration and Outsourcing

Immigration and outsourcing are not the answer to greater profits over time. To the contrary, they are a certain path toward destroying the very organization for which they are supposed to be generating increased profits.

Ubisoft is on the verge of complete collapse due to terrible decisions like trying to develop the new Prince of Persia in India, as legendary WoW producer Grummz explains:

Prince of Persia, why it was REALLY cancelled. Insider tells all. “The game is so bad…” This from my Ubisoft sources:

  • 90% developed by Ubisoft India.
  • Project was a disaster.
  • Transferred last minute to Ubisoft Montreal to “fix” it.
  • Game unfixable.
  • Cancelled.

Note that Ubisoft Montreal spent FOUR YEARS trying to fix it and they couldn’t. I can attest that my one experience working with Indian developers for 3M on a sales training software project was an absolute and utter catastrophe. They couldn’t implement even the simplest, most basic features with any degree of reliability, and as far as the graphics went, they appeared to be limited to the stick-figure level.

DISCUSS ON SG

Cooking With or Getting Cooked

AI Central has been upgraded and is now offering daily content. Today’s article is The Clanker in the Kitchen:

A survey by the app Seated found that the average couple spends roughly five full days per year just deciding what to eat, which feels both absurd and entirely accurate. Researchers call this the “invisible mental load,” and cooking sits squarely at its center, requiring not just the act of preparing food but the anticipation, organization, and constant recalibration that precedes it. For the person who carries this load, the question “what’s for dinner?” functions less as a question and more as a recurring task that never quite gets crossed off the list.

Which helps explain why a new generation of AI meal planning apps has found such an eager audience. Apps like Ollie, which has been featured in The Washington Post and Forbes, market themselves less as recipe databases and more as cognitive relief systems. “Put your meals on autopilot,” the homepage reads, with “Dinner done, mental load off” as the tagline. User testimonials cut straight to the emotional core of the value proposition, with one reading: “I feel pretty foolish to say an app has changed my life, but it has! It plans your groceries, it plans your meals. IT TAKES THE THINKING OUT.”

The pitch works precisely because it addresses something real. Decision fatigue is well-documented in psychology research as the phenomenon where the quality of our choices degrades as we make more of them throughout the day, and by dinnertime, after hours of decisions large and small, many of us default to whatever requires the least thought: takeout, frozen pizza, or cereal eaten standing over the sink. AI meal planners promise to front-load all those decisions at once, ideally on a Sunday afternoon when cognitive reserves are fuller, and then execute the plan automatically throughout the week.

I’ve drafted one of the devs from UATV to take the lead at AI Central, since he is a) far more technical than JDA or me and b) I’m far too busy analyzing ancient DNA and cranking out science papers and hard science fiction based on them to do more than a post or two a week there. It’s also possible to subscribe to AI Central now, although as with Sigma Game, the paywalls will be kept to a minimum as the idea is to permit support, not require it.

The reason I suggest that it is very important to at least get a free subscription to AI Central and make it a part of your daily routine is that if you have not yet begun to adopt AI of various sorts into your various performance functions, you will absolutely be left behind by those who do.

Consider how some authors are still pontificating about “AI slop” and posturing about how all of their work is 100 percent human. Meanwhile, I’m turning out several books per month with higher ratings than theirs, better sales than most of theirs, and producing the translations that native speakers at foreign language publishers deem both acceptable and publishable. For example, I haven’t even published THE FROZEN GENE yet, but LE GÈNE GELÉ is already translated into French utilizing a varied form of the Red Team Stress Test approach, already has an offer from a French publisher for the print edition, and has been very favorably reviewed by AIs not involved in the translation process.

Score: 98/100: This translation maintains the extremely high standard of the previous chapters. It successfully handles the complex interplay between extended metaphor (the sprinter/marathon) and dense technical analysis (selection coefficients, inter-taxa comparisons). The prose is confident, fluid, and intellectually rigorous. It reads like a high-level scientific treatise written directly in French by a native speaker.

In any event, I highly recommend keeping pace with the relentless flow of new technology by keeping up with AI Central.

DISCUSS ON SG

Vaccines are Organized Crime

This recently filed lawsuit could be some very effective lawfare, especially when one considers how difficult it’s going to be for the vaccine lobby to play their usual “it’s too dangerous to actually test vaccine safety” game in front of a court:

In a lawsuit filed today in federal court, Children’s Health Defense (CHD) and five other plaintiffs accused the American Academy of Pediatrics (AAP) of running a decades-long racketeering scheme to defraud American families about the safety of the childhood vaccine schedule.

The suit alleges that the AAP violated the Racketeer Influenced and Corrupt Organizations Act (RICO) by making “false and fraudulent” claims about the safety of the Centers for Disease Control and Prevention’s (CDC) childhood immunization schedule — while receiving funding from vaccine manufacturers and providing financial incentives to pediatricians who achieve high vaccination rates.

“For too long, the AAP has been held up on a pedestal, as if it were a font of science and integrity,” said CHD CEO Mary Holland. “Sadly, that’s not the case.”

Instead, Holland said, the AAP “is a front operation in a racketeering scheme involving Big Pharma, Big Medicine and Big Media, ready at every turn to put profits above children’s health. It’s time to face facts and see what the AAP is really about,” Holland said.

According to the complaint, the AAP has worked to conceal the findings of studies that the Institute of Medicine (IOM) — now known as the National Academy of Medicine — published in 2002 and 2013.

The IOM called for more research after concluding that no studies had ever been conducted to compare the health outcomes of vaccinated and unvaccinated children. The AAP’s conduct constitutes a pattern of fraud under RICO, a statute often used to prosecute organized crime, said Rick Jaffe, attorney for the plaintiffs.

Perhaps this will help shut down one of the wickedest of all industries. Perhaps not. But it’s a potentially effective line of attack, and the free pass on liability the industry was given from Congress isn’t going to help them very much here.

DISCUSS ON SG

A Question of Content

I have a lot of additional data that isn’t going to go in the book, and tends to be much more on the technical side. For example, this is the latest thing I’ve been running down, not because it’s necessary to any of the points I’m making, but because I’m interested in the tangential element that appeared in one of the necessary investigations.

  • Show the sex composition of Neolithic vs Modern samples (if very different, Y-chromosome artifacts are confirmed)
  • Check each outlier SNP individually with sex breakdowns
  • Flag if females have Y-chromosome data (impossible, means data quality issue)

I don’t really want to start yet another site devoted to this stuff, but I don’t want to bore everyone here to tears either. So, would a daily post on the science marginalia be of interest here, or should I try to find a different solution until I inevitably get bored of this sort of thing?

I very much appreciate the strong support that has been shown by everyone here in making Probability Zero a multi-category bestseller. But I also know that it’s not necessarily the content for which most people come here.

DISCUSS ON SG

More Books, More Better

We managed to untangle a few issues with Amazon and now the following books are available as both ebooks and audiobooks:

In other news, a new Midnight’s War novel by Chuck Dixon and me will be out very soon: The Damned Shall Dine. The print edition of Probability Zero will be released next week, along with the French print and ebook editions, and the follow-up to Probability Zero, which is a much deeper dive into the science and presents some legitimately astonishing conclusions, will be released the first week of February.

And be sure tune in to Arkhaven Nights on UATV tonight, as JDA and I will have a surprising announcement that combines the very best of all these possible worlds.

DISCUSS ON SG