Sorry for delayed response I have been riding trains inCentral and South Asia. I am emailing this response to you at the same time Iam posting it on Mifnet
My 19 Sep post was in the context of a Mifnet discussionthat began in July that had been provoked by industry press reports that GlenHauenstein had told investors that Delta expected major returns from itsongoing project to have “artificial intelligence” set prices. My original Julypost suggested that if Hauenstein’s claims could be confirmed this could be oneof the biggest stories in aviation, we need to figure out what’s actually goingon. The July discussion focused on three questions: 1. Was the Delta/Fletcherr “AI” project designed to giveDelta the ability to implement first-degree price discrimination (whether implemented now or in the future) so thatindividuals making identical pricing requests at the same point in time mightbe shown different prices based on Delta’ perceptions of their price elasticity? 2. What exactly did Hauenstein mean when he said AI wouldset Delta prices? What would Large Language Model/Generative AI pricing dodifferently than Delta’ longstanding PYM processes? Was Delta expecting “AI” tomake marginal improvements to its longstanding RM function or significantlytransform them? If the latter, what major RM shortcomings would “AI” solve?What totally new capabilities would “AI price setting” introduce? 3. In order of magnitude terms, were the revenue gains fromintroducing “AI pricing” likely to be incremental/useful or huge/competitivelypowerful? Was increased revenue from first-degree price discrimination animportant part of the reason for Hauenstein’s optimism with his investors?  Neither Delta’s public statements nor subsequent presscoverage provided clear answers to these questions. Subsequent Delta statementsimplied they don’t intend to introduce first-degree price discrimination, but theyhaven’t ruled it out in the future and have made no effort to demonstrate that thebig gains Hauenstein is promising investors come from sources other than first-degreeprice discrimination.  Again, the commentfrom Senator Gallego who thinks Hauenstein is either lying to his investors(about huge potential gains) or to Congress (about disavowing first-degreeprice discrimination). The subsequent Mifnet discussion recognized that any effortto implement first-degree price discrimination would be a huge industry issuebut would face a variety of challenges. While we could speculate, no one seemedable to confidently answer the three questions. Once Ronell distributed your paper, the question for theMifnet discussion was whether it helped answer the three questions. I stand bywhat I told the Mifnet on the 19th—no it didn’t, and I providedexplanations with direct quotes justifying that conclusion. Your response notonly failed to clarify how you would answer the three questions but ignoredalmost all the evidence I presented demonstrating that you didn’t have clear answersto those questions. Will the AI project enable first-degree pricediscrimination? Biggest question of all. We all know airlines have long hadaccess to lots of “personal information.” But is the Delta/Fletcherr projectdesigned to change the use this information so Delta can show different pricesto different customers? Your paper ignored this, and when I highlighted it inmy post of the 19th you ignored it again. Your paper misrepresentedpublic concerns about the use of anti-competitive market power in order tocapture consumer surplus as concerns about data privacy. Your response ignored my objection that thesingle quote from a Georgia TV station you used to represent the backlash toHauenstein’s statements similarly misrepresented the industry/public debateover first-degree price discrimination. Will AI drive major, dramatic changes to Delta RM, andif so how? No one could read your paper without missing your Very StrongImplications that this is a Really Big Deal. Your paper used the wordsrevolution/revolutionary ten times. Delta/Fletcherr “have begun to automateairline pricing.” As I said on the 19th, any RM function can beimproved, but your response ignored my stated concern that all of theimprovements your paper cited sounded like things airline RM has been doing foryears and fall well short of things one could call revolutionary. Delta airlinepricing isn’t currently automated? Airlines are incapable of implementing theirRM models? Airline data isn’t properly warehoused? I invited you to offerconcrete illustrations of major Delta RM shortcomings that the AI project couldfix; your reply failed to offer any, much less ones that could be consideredrevolutionary. Is Delta’s introduction of “AI pricing” something withthe potential to massively increase revenue and disrupt industry competition?No one could read your paper without missing your Very Strong Implications thatthis will have Really Huge Benefits.  Delta/Fletcherr “AI pricing” approach could increaserevenue by 10%, could be higher than 10%, and Delta’ competitors are ceding abig revenue advantage by failing to adopt the Delta/Fletcherr approach. You ignored my objection that this potential wasn’tqualified in any way and if taken at face value would produce the biggestprofit gain in industry history. You then insisted that the existence ofuncertainty meant you could ignore my objection that you hadn’t provided thedetailed concrete examples that a conclusion this powerful would normallywarrant or explained why major airlines like Delta were unable to achieve themin the past. The AI company claimsrevenue improvements of 10% due to pricing optimization from the model. We donot know if Delta is seeing the claimed 10% in their test cases, but we do knowthe airline is very, very satisfied with the results." You were the author of this paper. By endorsing Fletcherr’sspecific 10% claim and Delta’s very satisfied comment you are clearly tellingyour readers that you believe gains of this magnitude are highly plausible. Ofcourse Delta hasn’t publicly committed to specific improvement targets. Thatfact isn’t “adding context and nuance”. It is trying to avoid accountabilityfor the findings and opinions you’ve presented and trying to blow off requeststhat the findings and opinions you present be properly substantiated. Now if Delta/Fletcherr are really focused on creating afirst-degree price discrimination system perhaps adjectives like revolutionaryand claims of $5 billion annual revenue gains might become a bit more plausible.The Mifnet (and a great deal of other industry) discussion is trying to figureout if that’s what happening. But your paper steadfastly refused to address theissue and refused to offer any explanation of how revolutionary/multi-billion-dollarimprovements could be possible, if Delta was being fully honest when it toldCongress that it had no intention of introducing first-degree pricediscrimination. There was nothing in your response that claimed thatanything I’d said on the 19th was factually wrong or could bechallenged using public evidence. Instead it focused on a lot of emotionalpoints that frankly don’t make much sense. You were “disappointed” you hadn’t received my comments inadvance. Your published paper was not published at a discussion forum and didnot specifically invite comments. As you said, your paper was offered as acontribution to industry discussion and that’s how it was taken when Ronelldistributed it to Mifnet participants. Mifnet participants comment on lots ofexternal comments about industry issues. I recognize that you were not directly involved when theMifnet first discussed the three questions about the Delta/Fletchrr approach,but I don’t see how anyone could discuss anything related to airline AI pricingwithout considering them. You complained that your paper said it wasn’taddressed to experts in data science but I’m not aware of anyone on the Mifnetwho is on the cutting edge of data science and nothing in my post involved datascience issues. Your complaint of objections coming from people who conflate“artificial intelligence” with ChatGPT assisted internet searches is strawmanningand a bit insulting. You claimed my post was “angry”,“disingenuous” “appears to have been taken personally, somehow. Perhaps someone forwarded the paper to theposter with a nasty note?” and was an attack on your integrity. None of theseemotional comments referred to any specific thing I’d written and I found theseaspects of your reply unwarranted and totally unprofessional.
-------------------------------------------------------------------------- Revised: 20250507 You are receiving The Mifnet because you requested to join this list. The Mifnet is largely a labor of love, however the infrastructure isn't exactly cost-free. If you'd care to make a small contribution to the effort, please know that it would be greatly appreciated: https://wardell.us/url/mifbit All posts sent to the list should abide by these policies: 1) List members acknowledge that participation in Mifnet is a privilege--not a right. 2) Posts are always off the record, absent specific permission from the author. 3) The tone of discussions is collegial. 4) Posts are expected to be in reasonably good taste. 5) We discuss ideas and not personalities, and we don't speak ill of other Mifnet members. * The Mifnet WEB SITE is: https://www.mifnet.com/ * To UNSUBSCRIBE from this list at any time please visit: https://lists.mifnet.com/ OR: SEND THIS MESSAGE via email: mifnet-requ...@lists.mifnet.com?subject=leave * Send Mifnet mailing list POSTS/SUBMISSIONS to: mifnet@lists.mifnet.com * You may reach the person managing The Mifnet at: mifnet-ow...@lists.mifnet.com * Please consider the DIGEST version of The Mifnet, which consolidates all list traffic into 1-3 messages daily. See instructions at: https://lists.mifnet.com/ * Manage your personal Mifnet SUBSCRIPTION at: https://lists.mifnet.com/ * For a list of all available Mifnet commands, SEND THIS MESSAGE via email: mifnet-requ...@lists.mifnet.com?subject=help * View The Mifnet LIST POLICIES and PRIVACY POLICY at: https://mifnet.com/index.php/policies * View instructions for Mifnet DELIVERY PROBLEMS at: https://mifnet.com/index.php/delivery-problems * View The Mifnet LIST ARCHIVE at: https://lists.mifnet.com/hyperkitty/list/mifnet@lists.mifnet.com/
