Hey, We Know This Is Fake News: Sharing Secrets on Signal

April 24, 2025

Some government professionals allegedly blundered when they accidentally shared secret information via Signal with a reporter. The reporter, by the way, is not a person who wears a fedora with a command on it. To some, sharing close-hold information is an oopsie, but doing so with a non-hat wearing reporter is special. The BBC explained what the fallout will be from this mistake: “Why Is It A Problem If Yemen Strike Plans Shared On Signal?”

The Atlantic editor-in-chief Jeffrey Goldberg alleged received close-hold information via a free messaging application. A government professional seemed to agree that the messages appeared to be authentic. Hey, a free application is a cost reducer. Plus, Signal is believed to be encrypted end to end and super secure to boot. Signal is believed to be the “whisper network” of Washington D., an area known for its appropriate behavior and penchant for secrecy. (What was Wilbur Mills doing in the reflecting pool?)

While the messages are encrypted, bad actors (particularly those who may or may not be pals of the United States) allegedly can penetrate the Signal system. The Google Threat Intelligence Group noticed that Russia’s intelligence services have stepped up their hacking activities. Well, maybe or maybe not. Google is the leader in online advertising, but its “cyber security” expertise was acquired and may not be Googley yet.

The US government is not encouraging use of free messaging apps for sensitive information. That’s good. And the Pentagon is not too keen on a system not authorized to transmit non-public Department of Defense information. That’s good to know.

The whole sharing thing presents a potential downside for whomever is responsible for the misstep. The article says:

“Sensitive government communications are required to take place in a sealed-off room called a Sensitive Compartmentalized Information Facility (SCIF), where mobile phones are generally forbidden. The US government has other systems in place to communicate classified information, including the Joint Worldwide Intelligence Communications System (JWICS) and the Secret Internet Protocol Router (SIPR) network, which top government officials can access via specifically configured laptops and phones.”

Will there be consequences?

The article points out that the slip betwixt the cup and the lip may have violated two Federal laws:

“If confirmed, that would raise questions about two federal laws that require the preservation of government records: the Presidential Records Act and the Federal Records Act. "The law requires that electronic messages that take place on a non-official account are preserved, in some fashion, on an official electronic record keeping system," said Jason R Baron, a former director of litigation at the National Archives and Records Administration. Such regulations would cover Signal, he said.”

Hey, wasn’t the National Archive the agent interrupting normal business and holiday activities at a high profile resort residence in Florida recently? What does that outfit know about the best way to share sensitive information?

Whitney Grace, April 24, 2025

Israel Military: An Alleged Lapse via the Cloud

April 23, 2025

dino orange_thumbNo AI, just a dinobaby watching the world respond to the tech bros.

Israel is one of the countries producing a range of intelware and policeware products. These have been adopted in a number of countries. Security-related issues involving software and systems in the country are on my radar. I noted the write up “Israeli Air Force Pilots Exposed Classified Information, Including Preparations for Striking Iran.” I do not know if the write up is accurate. My attempts to verify did not produce results which made me confident about the accuracy of the Haaretz article. Based on the write up, the key points seem to be:

  1. Another security lapse, possibly more severe than that which contributed to the October 2023 matter
  2. Classified information was uploaded to a cloud service, possibly Click Portal, associated with Microsoft’s Azure and the SharePoint content management system. Haaretz asserts: “… it [MSFT Azure SharePoint Click Portal] enables users to hold video calls and chats, create documents using Office applications, and share files.”
  3. Documents were possibly scanned using CamScanner, A Chinese mobile app rolled out in 2010. The app is available from the Russian version of the Apple App Store. A CamScanner app is available from the Google Play Store; however, I elected to not download the app.

image

Modern interfaces can confuse users. Lack of training rigor and dashboards can create a security problem for many users. Thanks, Open AI, good enough.

Haaretz’s story presents this information:

Officials from the IDF’s Information Security Department were always aware of this risk, and require users to sign a statement that they adhere to information security guidelines. This declaration did not prevent some users from ignoring the guidelines. For example, any user could easily find documents uploaded by members of the Air Force’s elite Squadron 69.

Regarding the China-linked CamScanner software, Haaretz offers this information:

… several files that were uploaded to the system had been scanned using CamScanner. These included a duty roster and biannual training schedules, two classified training presentations outlining methods for dealing with enemy weaponry, and even training materials for operating classified weapons systems.

Regarding security procedures, Haaretz states:

According to standard IDF regulations, even discussing classified matters near mobile phones is prohibited, due to concerns about eavesdropping. Scanning such materials using a phone is, all the more so, strictly forbidden…According to the Click Portal usage guidelines, only unclassified files can be uploaded to the system. This is the lowest level of classification, followed by restricted, confidential, secret and top secret classifications.

The military unit involved was allegedly Squadron 69 which could be the General Staff Reconnaissance Unit. The group might be involved in war planning and fighting against the adversaries of Israel. Haaretz asserts that other units’ sensitive information was exposed within the MSFT Azure SharePoint Click Portal system.

Several observations seem to be warranted:

  1. Overly complicated systems involving multiple products increase the likelihood of access control issues. Either operators are not well trained or the interfaces and options confuse an operator so errors result
  2. The training of those involved in sensitive information access and handling has to be made more rigorous despite the tendency to “go through the motions” and move on in many professionals undergoing specialized instruction
  3. The “brand” of Israel’s security systems and procedures has taken another hit with the allegations spelled out by Haaretz. October 2023 and now Squadron 69. This raises the question, “What else is not buttoned up and ready for inspection in the Israel security sector?

Net net: I don’t want to accept this write up as 100 percent accurate. I don’t want to point the finger of blame at any one individual, government entity, or commercial enterprise. But security issues and Microsoft seem to be similar to ham and eggs and peanut butter and jelly from this dinobaby’s point of view.

Stephen E Arnold, April 23, 2025

Microsoft and Its Modern Management Method: Waffling

April 23, 2025

dino orange_thumb_thumb_thumb_thumbNo AI, just the dinobaby himself.

The Harvard Business School (which I assume will remain open for “business”) has not addressed its case writers to focus on Microsoft’s modern management method. To me, changing direction is not a pivot; it is a variant of waffling. “Waffling” means saying one thing like “We love OpenAI.” Then hiring people who don’t love OpenAI and cutting deals with other AI outfits. The whipped cream on the waffle is killing off investments in data centers.

If you are not following this, think of the old song “The first time is the last time,” and you might get a sense of the confusion that results from changes in strategic and tactical direction. You may find this GenX, Y and Z approach just fine. I think it is a hoot.

PC Gamer, definitely not the Harvard Business Review, tackles one example of Microsoft’s waffling in “Microsoft Pulls Out of Two Big Data Centre Deals Because It Reportedly Doesn’t Want to Support More OpenAI Training Workloads.”

The write up says:

Microsoft has pulled out of deals to lease its data centres for additional training of OpenAI’s language model ChatGPT. This news seems surprising given the perceived popularity of the model, but the field of AI technology is a contentious one, for a lot of good reasons. The combination of high running cost, relatively low returns, and increasing competition—plus working on it’s own sickening AI-made Quake 2 demo—have proven enough reason for Microsoft to bow out of two gigawatt worth of projects across the US and Europe.

I love the scholarly “sickening.” Listen up, HBR editors. That’s a management term for 2025.

The article adds:

Microsoft, as well as its investors, have witnessed this relatively slow payoff alongside the rise of competitor models such as China’s Deepseek.

Yep, “payoff.” The Harvard Business School’s professors are probably not familiar with the concept of a payoff.

The news report points out that Microsoft is definitely, 100 percent going to spend $80 billion on infrastructure in 2025. With eight months left in the year, the Softies have to get in gear. The Google is spending as well. The other big time high tech AI juggernauts are also spending.

Will these investments payoff? Sure. Accountants and chief financial officers learn how to perform number magic. Guess where? Schools like the HBS. Don’t waffle. Go to class. Learn and then implement big time waffling.

Stephen E Arnold, April 23, 2025

Management Challenges in Russian IT Outfits

April 23, 2025

dino orange_thumb_thumb_thumb_thumbBelieve it or not, no smart software. Just a dumb and skeptical dinobaby.

Don’t ask me how, but I stumbled upon a Web site called PCNews.ru. I was curious, so fired up the ever-reliable Google Translate and checked out what “news” about “PCs” meant to the Web site creator. One article surprised me. If I reproduce the Russian title it will be garbled by the truly remarkable WordPress system I have been using since 2008. The title of this article in English courtesy of the outfit that makes services available for free is, “Systemic Absurdity: How Bureaucracy and Algorithms Replace Meaning.”

One thing surprised me. The author was definitely annoyed by bureaucracy. He offers some interesting examples. I can’t use these in my lectures, but I found sufficiently different to warrant my writing this blog post.

Here are three examples:

  1. “Bureaucracy is the triumph of reason, where KPIs are becoming a new religion. According to Harvard Business Review (2021), 73% of employees do not see the connection between their actions and the company’s mission.”
  2. 41 percent of the time “military personnel in the EU is spent on complying with regulations”
  3. “In 45% of US hospitals, diagnoses are deliberately complicated (JAMA Internal Medicine, 2022)”

Sporty examples indeed.

The author seems conversant with American blue chip consultant outputs; for example, and I quote:

  1. 42% of employees who regularly help others face a negative performance evaluation due to "distraction from core tasks". Harvard Business Review (2022)
  2. 82% of managers believe cross-functional collaboration is risky (Deloitte, Global Human Capital Trends special report 2021).
  3. 61% of managers believe that cross-functional assistance “reduces personal productivity.” “The Collaboration Paradox” Deloitte (2021)

Where is the author going with his anti-bureaucracy approach? Here’s a clue:

I once completed training under the MS program and even thought about getting certified? Do they teach anything special there and do they give anything that is not in the documentation on the vendor’s website/books/Internet? No.

I think this means that training and acquiring certifications is another bureaucratic process disconnected from technical performance.

The author then brings up the issue of competence versus appearance. He writes or quotes (I can’t tell which):

"A study by Hamermesh and Park (2011) showed that attractive people earn on average 10-15% more than their less attractive colleagues. The work of Timasin et al. (2017) found that candidates with an attractive appearance are 30% more likely to receive job offers, all other things being equal. In a study by Harvard Business Review (2019), managers were more likely to recommend promotion to employees with a "successful appearance", associating them with leadership qualities"

The essay appears to be heading toward a conclusion about technical management, qualifications, and work. The author identifies “remedies” to these issues associated with technical work in an organization. The fixes include:

  1. Meta regulations; that is, rules for creating rules
  2. Qualitative, not just quantitative, assessments of an individual’s performance
  3. Turquoise Organizations

This phrase refers to an approach to management which emphasizes self management and an organic approach to changing an organization and its processes.

The write up is interesting because it suggests that the use of a rigid bureaucracy, smart software, and lots of people produces sub optimal performance. I would hazard a guess that the author feels as though his/her work has not been valued highly. My hunch is that the inclusion of the “be good looking to get promoted” suggests the author is unlikely to be retained to participate in Fashion Week in Paris.

An annoyed IT person, regardless of country and citizenship, can be a frisky critter if not managed effectively. I wonder if the redactions in the documents submitted by Meta were the work of a happy camper or an annoyed one? With Google layoffs, will some of these capable individuals harbor a grudge and take some unexpected decisions about their experiences.

Interesting write up. Amazing how much US management consulting jibber jab the author reads and recycles.

Stephen E Arnold, April 23, 2025

Honesty and Integrity? Are You Kidding Me?

April 23, 2025

dino orange_thumb_thumbNo AI, just the dinobaby himself.

I read a blog post which begins with a commercial and self promotion. That allowed me to jump to the actual write up which contains a couple of interesting comments. The write up is about hiring a programmer, coder, or developer right now.

The write up is “Tech Hiring: Is This an Inflection Point?” The answer is, “Yes.” Okay, now what is the interesting part of the article? The author identifies methods of “hiring” which includes interviewing and determining expertise which no longer work.

These methods are:

  1. Coding challenges done at home
  2. Exercises done remotely
  3. Posting jobs on LinkedIn.

Why don’t these methods work?

The answer is, “Job applicants doing anything remotely and under self-supervision cheat. Okay, that explains the words “honesty” and “integrity” in the headline to my blog post.

It does not take a rocket scientist or a person who gives one lecture a year to figure out what works. In case you are wondering, the article says, “Real person interviews.” Okay, I understand. That’s the way getting a job worked before the remote working, Zoom interviews, and AI revolutions took place. Also, one must not forget Covid. Okay, I remember. I did not catch Covid, and I did not change anything about my work routine or daily life. But I did wear a quite nifty super duper mask to demonstrate my concern for others. (Keep in mind that I used to work at Halliburton Nuclear, and I am not sure social sensitivity was a must-have for that work.)

Several observations:

  1. Common sense is presented as a great insight. Sigh.
  2. Watching a live prospect do work yields high value information. But the observer must not doom scroll or watch TikToks in my opinion.
  3. Allowing the candidate to speak with other potential colleagues and getting direct feedback delivers another pick up truck of actionable information.

Now what’s the stand out observation in the self-promotional write up?

LinkedIn is losing value.

I find that interesting. I have noticed that the service seems to be struggling to generate interest and engagement. I don’t pay for LinkedIn. I am 80, and I don’t want to bond, interact, or share with individuals whom I will never meet in the short time I have left to bedevil readers of this Beyond Search post.

I think Microsoft is taking the same approach to LinkedIn that it has to the problem of security for its operating systems, the reliability of its updates, and the amazingly weird indifference to flaws in the cloud synchronization service.

That’s useful information. And, no, I won’t be attending the author’s one lecture a year, subscribing to his for fee newsletter, or listening to his podcast. Stating the obvious is not my cup of tea. But I liked the point about LinkedIn and the implications about honesty and integrity.

Stephen E Arnold, April 23, 2025

FOGINT: Pavel Durov Responds to His Problems without Crisis PR Inputs

April 22, 2025

fog from gifer 8AC8 small Before the US National Cyber Crime Conference, possibly significant news emerged about Telegram.

In my Telegram lecture at the NCCC 2025, I don’t talk about the psychological and financial impact Pavel Durov experienced as a result of his interaction with the French government. He was greeted at a Paris airport and detained. He talked with French officials. He hired Kaminski and his associates to represent him in the legal matter. Within a few weeks of his being confined to France, although not in St. Denis or Maison d’arrêt de Fleury-Mérogis, he seemed to be showing interest in providing some government authorities with Telegram “user names” or made-up handles and phone numbers (some real and some obtained through services providing temporary phone numbers). Mr. Durov, via indirect communication methods, seemed to indicate that he was doing what he had always done: Followed the rules of the jurisdictions in which Telegram operated. Then, without going to the French equivalent of a trial, Mr. Durov was allowed to return to Dubai, United Arab Emirates, one of the countries which has granted him citizenship.

Upon his return, he took some interesting actions:

  1. He began fiddling with some knobs and dials in order to get the TONcoin out of its doldrums
  2. He ramped up the marketing activities of the TON Foundation. (Please, keep in mind that Telegram and its proxy say they do not do marketing. The Telegram entities also do not do personnel or content control either.) Telegram and the Open Network Foundation (aka TON Foundation or ONF) are “there is no there there” or virtual outifts making do with coffee shop meetings, Telegram Messenger interactions, and maybe some face-to-face activity in a rented temporary space.)
  3. He started talking on CNN, for example.

From one of my team, I learned a couple of new “facts” this morning (April 22, 2025, 6 am US Eastern time). This individual reported:

Telegram has never shared private messages. The source is https://news.az/news/t-elegram-has-never-shared-private-messages-durov-says. Now there are two types of messages on Telegram Messenger. First, there are the default messages. These are encrypted from sender to the Telegram command-and-control system. There the messages are decrypted and processed. Then the messages are re-encrypted and forwarded to the recipient who opens them. The second type of message is the “secret” message which requires that the recipient be on Telegram when the sender creates and sends the message. That message is forwarded by the command-and-control system to the recipient. These “secret” messages remain encrypted for their digital journey. What Telegram shares with law enforcement is the IP address and the phone number of the sender. I will leave it to you to consider the “value” to Telegram’s command-and-control center of logging the metadata for messages and what happens when an investigator receives a bogus or temporary mobile number.,

A second item he shared with me this morning is that Pavel Durov continues to find himself at the edge of chaos with Russia. According to Meduza.io:

Russia slaps Telegram with another multi-million-ruble fine for refusing to take down ‘prohibited content’.

Several observations may be warranted:

  1. For a company that does not “do” marketing, Pavel Durov has been a busy and willing marketer on behalf of the Telegram operation. In fact, he seems to be sending out the message, “We cooperate with law enforcement.” Then he tries to make clear that he doesn’t reveal too much when Telegram cooperates with investigators.
  2. As a Russian citizen, Mr. Durov may want to [a] work with certain Russian officials to create the appearance of a problem with Russia in order to reduce the pressure from France or [b] Russia is acting independently to let Mr. Durov know he is a person of interest to the Kremlin. In short, which is it? Is Mr. Durov an asset for a certain country or is he a problem for a certain country in which some of his family and possibly some of his “core developers” reside? Uncomfortable either way I think.
  3. Telegram and its crypto play are not enjoying significant TONcoin upsides. The TONcoin is not the would-be high flier, Hamster Kombat home run it was prior to Mr. Durov’s arrest.

What’s clear is that Mr. Durov’s legal problems in France have been resolved. What is Mr. Durov’s relationship with the Russia’s government? Fuzzy stuff.

Net net: Mr. Durov’s recent actions appear to be signals that suggest Telegram is going to have to pull a rabbit from someone’s hoodie.

Stephen E Arnold, April 22, 2025

Smart Software Exploits Direct Tuition Payment. Sure, the Fraud Is Automated

April 22, 2025

dino orange_thumb_thumb_thumb_thumb_thumbNo AI, just the dinobaby himself.

The Voice of San Diego published “As Bot Students Continue to Flood In, Community Colleges Struggle to Respond.” The write up is one of those recipes that “real” news outfits provide to inform their readers about a crime. When I worked through the article, my reaction was, “The process California follows for community college student assistance is a big juicy sandwich on a picnic table in the park on a warm summer day.”

Will the insects flock to the sandwich?

Absolutely. Plus, telling the insects where the sandwich is and the basics of getting their mandibles on that sandwich does one thing: Provide an easy-to-follow set of instructions for a bad actor to follow.

The write up says:

Kevin Alston, a business professor who has taught at Southwestern for nearly 20 years, has stumbled across even more troubling incidents. During a prior semester, he actually called some of the students who were enrolled in his classes but had not submitted any classwork.  “One student said ‘I’m not in your class. I’m not even in the state of California anymore’” Alston recalled.  The student told him they had been enrolled in his class two years ago but had since moved on to a four-year university out of state.  “I said, ‘Oh, then the robots have grabbed your student ID and your name and re-enrolled you at Southwestern College. Now they’re collecting financial aid under your name,’” Alston said.

The opportunity for fraud is a result of certain rules and regulations that require that financial aid be paid directly to the “student.” Enroll as a fake student and get a chunk of money. The more fake students that apply and receive aid, the more money the fake students receive.

California appears to be taking steps to reduce the fraud.

Several observations:

  1. A basket of rules and regulations appear to create this fraud opportunity
  2. Smart software in the hands of clever individuals allows the bad actors to collect money. (I am not sure how one cashes multiple checks made out to a fake person, but obviously there are ways around this problem. Are those nifty automatic teller machine deposits an issue?)
  3. The problem, according to the write up, has been known and getting larger since 2021.

I must admit that I think about online fraud in the hands of pig butchering outfits in the Golden Triangle. The fake student scam sounds like a smaller scale operation. Making a teacher the one who must identify the fake student does not seem to be working.

Okay, let’s see what the great state of California does to resolve this problem. Perhaps the instructors need to attend online classes in fraud detection, apply for financial aid, and get an extra benefit for this non-teaching work? Will community college teachers make good cyber investigators? Sure, especially those teaching history, social science, and literature classes.

Stephen E Arnold, April 22, 2025

JudyRecords: Is It Back or Did It Never Go Away?

April 22, 2025

dino orange_thumb_thumb_thumbBelieve it or not, no smart software. Just a dumb and skeptical dinobaby.

I was delighted to see that JudyRecords is back online. Here’s what the service says as of April 19, 2025:

Judyrecords is a 100% free nationwide search engine that lets you instantly search hundreds of millions of United States court cases and lawsuits.judyrecords has over 100x more cases than Google Scholar and 10x more cases than PACER, the official case management system of the United States federal judiciary.As of Jul 2022, judyrecords now features free full-text search of all United States patents from 1/1/1976 to 07/01/2022 — over 8.1 million patents in total.

My thought is that lawyers, law students, and dinobabies like me will find the service quite useful.

The JudyRecords’ Web site adds:

The first 500K results are displayed instead of just the first 2K.

  • murder – 926K cases
  • fraud – 2.1 million cases
  • burglary – 3.7 million cases
  • assault – 8.2 million cases

Most people don’t realize that the other “free” search engines limit the number of hits shown to the user. The old-fashioned ideas of precision and recall are not operative with most of the people whom I encounter. At the Googleplex, precision and recall are treated like a snappy joke when the Sundar & Prabhakar Comedy Show appears in a major venue like courtrooms.

If you want to control the results, JudyRecords provides old-fashioned and definitely unpopular methods such as Boolean logic. I can visualize the GenZs rolling their eyes and mouthing, “Are you crazy, old man?”

Please, check out JudyRecords because the outstanding management visionaries at LexisNexis, Thomson Reuters, and other “professional” publishers will be taking a look themselves.

Stephen E Arnold, April 22, 2025

Kiddie Loving Google and Data Hoovering

April 22, 2025

If you do not have kids or grandkids in school, you may have missed Google’s very successful foray into K-12 education. Google’s “Workspace for Education” tools are free to schools, but is the company providing them purely from a sense of civic duty? Of course not. Bloomberg Law reports, “Google Hit with Lawsuit over Data Collection on School Kids.” Apparently, US schools did not learn from Denmark’s 2022 ban on Google Workspace in its schools. Or they decided savings and convenience trumped student privacy and parental consent. Writer Isaiah Poritz tells us:

“Google LLC is unlawfully using its products—ubiquitous in K-12 education—to secretly gather information about school age children, substituting the consent of the school for that of parents, a proposed class action filed in California federal court said Monday. The tech giant collects not only traditional education records ‘but thousands of data points that span a child’s life,’ and ‘neither students nor their parents have agreed to this arrangement, according to the US District Court for the Northern District of California complaint.”

This is a significant breach, if true, considering almost 70% of K-12 schools in the US use these tools. We also learn:

“The company doesn’t disclose that it embeds hidden tracking technology in its Chrome browser that creates a child’s unique digital ‘fingerprint,’ the plaintiffs said. The fingerprint allows Google to ‘to track a child even when she or her school administrator has disabled cookies or is using technologies designed to block third-party cookies.’ The suit said Google has failed to obtain parental consent to take school childrens’ personal data. ‘Instead, Google relies on the consent of school personnel alone,’ the complaint said. ‘But school personnel do not have authority to provide consent in lieu of parents.’”

No, they do not. Or they shouldn’t. It seems like parents’ rights groups should have something to say about this. Perhaps they are too busy policing library shelves. The suit alleges Google is both selling students’ data to third parties and using it for its own targeted advertising. We note it would also be very easy, if the firm is so inclined, to build up a profile of a student who later creates a Google account which is then mapped onto that childhood data.

Naturally, Google denies the suit’s allegations. Of course, our favorite company does.

Cynthia Murrell, April 22, 2025

ArXiv: Will Other Smart Software Systems Get “Free” Access? Yeah, Sure

April 21, 2025

dino orangeBelieve it or not, no smart software. Just a dumb and skeptical dinobaby.

Before commenting on Cornell University’s apparent shift  of the ArXiv service to the Google Cloud, let me point you to this page:

image

The page was updated 15 years ago. Now check out the access to

NCSTRL, the Networked Computer Science Technical Reference Library.

CoRR, the Computing Research Repository.

The Open Archives Initiative.

ETRDL, the ERCIM Technical Reference Digital Library.

Cornell University Library Historical Math Book Collection

Cornell University Library Making of America Collection

Hein online Retrospective Law Journals

Yep, 404s, some content behind paywalls, and other data just disappeared because Bing, Google, and Yandex don’t index certain information no matter what people believe or the marketers say.

This orphaned Cornell University Dienst service has “gorged out”; that is, jumped off a bridge to the rocks below. The act is something students know about but the admissions department seems to not be aware of the bound phrase.

I read “Careers at ArXiv.” The post seems to say to me, “We are moving the ArXiv “gray” papers to Google Cloud. Here’s a snippet of the “career” advertisement / news announcement:

We are already underway on the arXiv CE ("Cloud Edition") project. This is a project to re-home all arXiv services from VMs at Cornell to a cloud provider (Google Cloud). There are a number of reasons for this transition, including improving arXiv’s scalability while modernizing our infrastructure. This will not be a simple port of the existing arXiv code base because this project will:

  • replace the portion of our backends still written in perl and PHP
  • re-architect our article processing to be fully asynchronous, and provide better insight into the processing workflows
  • containerize all, or nearly all arXiv services so we can deploy via Kubernetes or services like Google Cloud Run
  • improve our monitoring and logging facilities so we can more quickly identify and manage production issues with arxiv.org
  • create a robust CI/CD pipeline to give us more confidence that changes we deploy will not cause services to regress

The cloud transition is a pre-requisite to modernizing arXiv as a service. The modernization will enable: – arXiv to expand the subject areas that we cover – improve the metadata we collect and make available for articles, adding fields that the research community has requested such as funder identification – deal with the problem of ambiguous author identities – improve accessibility to support users with impairments, particularly visual impairments – improve usability for the entire arXiv community.

I know Google is into “free.” The company is giving college students its quantumly supreme smart software for absolutely nothing. Maybe a Google account will be required? Maybe the Chrome browser may be needed to give those knowledge hungry college students the best experience possible? Maybe Google’s beacons, bugs, and cookies will be the students’ constant companions? Yeah, maybe.

But will ArXiv exist in the future? Will Google’s hungry knowledge munchers chew through the data and then pull a Dienst maneuver?

As a dinobaby, I liked the ArXiv service, but I also liked the Dienst math repository before it became unfindable.

It seems to me that Cornell University is:

  1. Saving money at the library and maybe the Theory Center
  2. Avoiding future legal dust ups about access to content which to some government professionals may reveal information to America’s adversaries
  3. Intentionally or inadvertently giving the Google control over knowledge flow related to matters of technical and competitive interest to everyone’s favorite online advertising company
  4. Running a variation of its Dienst game plan.

But I am a dinobaby, and I know zero about Cornell other than the “gorging out” approach to termination. I know even less about the blue chip consulting type thinking in which the Google engages. I don’t even know if I agree that Google’s recent court loss is really a “win” for the Google.

But the future of the ArXiv? Hey, where is that bridge? Do some students jump, fall, or get pushed to their death on the rocks below?

PS. In case your German is rusty “dienst” means duty and possibly “a position of authority” like a leader at Google.

Stephen E Arnold, April xx, 2025

« Previous PageNext Page »

  • Archives

  • Recent Posts

  • Meta