Archive for Anthropic

AI in a storm [cover]

Posted in Books, pictures, University life with tags , , , , , , , , , , , , , , , , , , on October 6, 2026 by xi'an

Tidy up, Claude!

Posted in Linux, Mountains, pictures, Statistics, Travel, University life with tags , , , , , , , , , , , , , , , , , on September 27, 2026 by xi'an

Nothing out of the ordinary there, but after years of procrastination at failing to update my publication webpage, following a wreckage by the default web composer on Ubuntu (and yours truly obviously!), I took advantage of the extra energy gained by my altitude night at the 5th station to ask Claude to update this page, with it replying it preferred to start from scratch (and from my LaTeX vita files). The only drawback is that many of the links introduced in the earlier version were not retrieved by the AI, despite my repeated request. On the plus side, trawling for missing entries and unpublished arXiv oldies but goldies, as well as adding mountain pictures (from the ‘Og!) proved much easier than the manual-by-keyboard approach. And I got an extra bonus in updating my LaTeX publication file and by creating the corresponding bibTeX file, all now available on my github page.

“too dangerous to release” [a correct prediction]

Posted in Books, Travel with tags , , , , , , , , , , , , on July 9, 2026 by xi'an

Nature of 26 May had a long report on how the Mythos model is deemed to be too dangerous by its conceptor, Anthropic, to be publicly released, but instead shared with a limited number of organisations of their own choice. The argument being that Mythos was able to spot security flaws in every operating system and web browser on the market. But, sooner or later, the AI will prove accessible to a hacker who manages to penetrate one of these organisations, won’t it?! As pointed out by Nature’s writers, the natural question is why should firms be left to decide on who gets access to their software and how dangerous they are. The same writers also muse about governments imposing prohibitions on the firms, including outside their country. Which is exactly what happened a few weeks later on 12 June with the Trump administration “suspend[ing] all access to both Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States” with little factual argumentation. Another illustration of the short-sighted, knee-jerk, inconsistent, policies of that joke of an administration that does not address the real issue.

class action, really?!

Posted in Books, Travel, University life with tags , , , , , , , , , , , , , , , on January 13, 2026 by xi'an

The letter from the US I found in my (home) mailbox last week sounded like a spam building on copyright violations, if an elaborate one, with promises of huge settlement benefits ($3000 per work, for a total of $1.5 billion!) and a California judge authorization stamped on the envelope… However, after being contacted by a co-author and checking on Internet about the existence of a class action against Anthropic and its pirated libraries, I realised this was not (a spam) and checked that the database of the works concerned by this settlement included eight of my books. (Incl. second editions.) Although I do not expect much return (if any!) once the costs and fees and publishers’ share are subtracted, and the remainder split between 7M books!, this is a first instance of getting back at the providers of pirated copies that are everywhere (since publishers came up with the brilliant scheme of provided access to pdf versions!)

“The lawsuit alleges that Anthropic infringed copyrights by downloading datasets containing copyrighted books in violation of the federal Copyright Act. Anthropic denies all the allegations and denies that it did anything wrong. Anthropic argues that its use of the downloaded datasets was fair use. You can get more information about the lawsuit and view related court documents (…) . Copying a work without permission is not copyright infringement if a defendant can show the copying was fair use. If the use is determined to be infringement, the Copyright Act provides for statutory damages of between $200 and $150,000 per work, depending on factors including the harm that was actually caused by the infringement, and whether the alleged infringer reasonably believed its use was fair or instead acted wilfully. If the use was fair (or there was no copying), the defendant owes $0 (…) The resulting Settlement is the largest copyright class action settlement in history. It provides approximately $3,000 per work (not per Class Member), plus interest earned on the Settlement fund, less the Court-approved costs and fees taken out“