It’s been a rough few weeks


But we’re back online again. First we had Bluehost arbitrarily decide to cut off the DNS, and the latest problem was a fart in the php that killed our cloudflare service. This stuff is all essential for functioning online, because the internet is a savage, ruthless environment. I remember when it was all just me running a blog on a computer in my lab with Apache doing all the hard work — I sometimes looked on incoming traffic, and there was a stunning quantity of spam and malignant stabs at the server every second, all blocked by Apache, and now the threats out there are even more common and more subtle. We’re working for now, though, until the next disaster.

There are worse disasters, though. To put it into context, watch this lethal level of sarcasm.

I won’t say it could be worse…rather, it will get worse.

Comments

  1. raven says

    …and now the threats out there are even more common and more subtle.
    and
    I won’t say it could be worse…rather, it will get worse.

    I know all that first hand.

    In July, an AI powered malware attack on my computer resulted in something getting into my bank account in about 15 minutes. I immediately shut down my bank account and 401(K) account, which it was also trying to get into.

    I nuked my credit card and checking account, froze my 401(K) account, and spent 3 days digging remote control and anti-antiviral programs out of my computer.
    No one lost any money but it wasn’t for someone and something’s lack of trying.

    Then, two weeks ago on a popular Hollywood website (tmz . cm, don’t go there), the same thing happened. The details were different but the sequence was the same.
    I think these AI malware programs are coming through in the ad channels.
    Since I’d already seen it once, I knew what to do and shut it down immediately before anything got downloaded.

    I had to notify a lot of people about the attack, and change all my financial accounts.
    People were sceptical at the time, when I mentioned “AI attack”.
    Now I’m getting emails from all over…warning me about AI attacks.

  2. robro says

    Good to have you back, PZ. Hopefully the ghost of Ed will fade away now, except in memories.

    raven @ #1 — “an AI powered malware attack on my computer resulted in something getting into my bank account in about 15 minutes.” If I may ask, what tipped you off? Was there a specific symptom? And what made you think it was “AI powered”?

  3. raven says

    @robro

    I lost control of my computer.
    It froze up and then a remote control program downloaded and took control.
    I just had to watch it get into my bank account and transfer out $24,000.
    I called the bank 800 number and told them to immediately freeze my account , which they did.
    They also said the transfer out wasn’t real and they would have caught it.
    They had no problem setting up a new account and sending me a new credit card.

    I also now have fraud warnings and freezes on the credit reporting companies like Equifax and Transunion.

    It just happened really fast.
    Whatever it was also disabled all of my antiviral software and turned off all my firewalls.

    I looked this type of attack up on Google and the security websites all said AI assisted scam.

  4. cheerfulcharlie says

    I just voted. The way it works in Houston, Texas, is on votes on an electric system that then prints out a filled ballot that gets turned in for tabulating, This leaves a paper trail, beloved of MAGA types. Naturally, the printer died. So I had to wait until somebody got the printer working. Which they did so I have officially voted. Hock patooie to Paxton and Abbot

  5. Hemidactylus says

    Always good to see the blog come back to life.

    I think there was at least one if not two snippets from Reed Timmer’s channel in that faux Exxon video. Even if brief his hyperbolic drama voice is instantly recognizable.

  6. says

    I thank PZ for spending all the time and hassle to get pharyngula back on line AGAIN.
    Sadly, PZ has had to deal with these two abusive corporations:
    » cloudflare, on trust pilot, has a rating of 1.2 out of 5.
    » bluehost is condemned by hundreds of webmasters as crap.
    Yes, there are too many thousands of malicious attacks on websites.

      AI bots are scraping terrabytes of data from millions of sites, bogging them down or crashing them.
      Cloudflare SUCKS and is NOT helpful. They are used by many sites as a middleman to ‘check if you are a bot’. But, if you are not using the latest version of a VERY few browsers they accept, or if your computer clock is set to the wrong time zone, they will prevent you from getting to the intended site.
      I ran duckduckgo (wouldn’t touch g00gle with a 10ft cattle prod) and found procedures for migrating your wordpress powered site from blue host to a better one. BUT, it is a complex and difficult procress. Our inhouse webmaster said they are not ‘up to date’ enough to accomplish that without many hours of studying. There might be a decent hosting company that would do that. But, I don’t know which one or the cost. Our hosting since 2003 has been excellent: qwknet. They have tools to run wordpress, but I don’t think they have someone to help you migrate your site.

    Our founder has held free computer clinics for years. Starting in ~2003 he has always warned people new to the internet:
    “The internet is like a being stranger being dumped in NYC. If you turn one way, they are in a brightly lit street with nice stores. If you turn the wrong way, you’re in a dark alley being mugged!”
      However, it’s not hyperbole saying that we are all in danger. miscreants can and will ruin your life by cracking your bank and healthcare data even if your computer isn’t hijacked.

  7. says

    Raven, as scary as that all was, it was probably the setup. That your bank told you the $24k alleged transfer was “fake” is a tipoff in itself. The next step would have been to send you fake alerts from financial institutions trying to have you call/text/click a link/send an e-mail to a fake number/address, at which time the real fun starts. This scheme is evolution-to-the-‘net/phone of the schemes developed when over-the-phone banking became a thing a few decades back.

    Always, always, always keep a carefully-proofread, offline and preferably printed but in a secure place, list of the correct contact information for every important account. Including your utilities, your ‘net service, and (naturally enough!) your cell carrier. When you spot suspicious activity, that’s what you refer to. (If you can memorize all of your account numbers, so much the better.) If you’re under attack, don’t rely on the system that’s under attack, even if you’re not doing the “naïve enough to click the link” thing.

  8. says

    OH, CRAP! I just watched the youtube PZ embedded. We need to remember that these FOSSIL FUEL FUCKHEADS ARE THE ENEMY OF PLANET EARTH. THE ‘EXXON VALDEZ’ WAS JUST AN APPETIZER. Now, I’m hearing the Animals song from the 1960’s “we gotta get outta this place, if it’s the last thing we ever do”. BUT, then Martha and the Vandllas warn us: “nowhere to run to, baby. Nowhere to hide” And, all this while tens of thousands of people can’t hear it over the noise and pollution of the data centers just a few feet from their houses!
    I don’t have the resources to find a safe place to live.
    I don’t think any decent, caring person could disagree with that.

  9. says

    I know that Raven is careful. But, @9 Jaws provided very good advice. Never contact your banksters or healthcare through a link or E-mail.
    I admonish everyone, don’t click on ads and be very careful with QR codes, they are blind links to who knows where.
    The printed list Jaws recommends is very important and is what we do.

  10. John Morales says

    Ridiculous hating on Cloudflare, shermanj. Quite ignorant.

    It is quite reputable, in fact, “According to W3Techs, Cloudflare is used by approximately 21.3% of all websites on the Internet as of January 2026.” (Wikipedia).

  11. robro says

    shermaj @ #11 – I don’t contact any business, government agency (e.g. Social Security), or even friends by clicking links in emails. I get tons of emails “from” friends all the time with links to pictures they want to share with me. I even get those emails from myself…at least it’s my name in the from field…god knows what the actual email address is for any of these.

  12. says

    @12 John Morales wrote: Ridiculous hating on Cloudflare, shermanj. Quite ignorant. It is quite reputable.. .
    I reply: John, are you a webmaster?
    You can try to criticize the factual reality we observe first hand if you want. You are making a false accusation. Our webmaster and I are reporting factually, accurately and disdainfully, not hatefully, from a perspective of decades of knowledge and experience in the field, along with credible info from other webmasters.

    If cloudflare is so great, why have they screwed PZ and so many others so many times?
    I don’t want to play your fatuous games anymore.

  13. John Morales says

    You are the one playing games, shermanj.

    You cherry-picked one rating, I didn’t.

    Cloudflare has a rating of 1.4 out of 5 on Trustpilot based on reviews from everyday internet users and end-consumers.
    Cloudflare has a rating of 4.5 out of 5 on G2 based on reviews from B2B software buyers and IT professionals.
    (https://www.g2.com/products/cloudflare-application-security-and-performance/reviews)
    Cloudflare has a rating of 4.7 out of 5 on Capterra based on reviews from small to medium business software buyers.
    (https://www.capterra.com/p/155191/Cloudflare/reviews/)
    Cloudflare has a rating of 4.5 out of 5 on Gartner Peer Insights based on reviews from enterprise-level IT leaders and network engineers.
    (https://www.gartner.com/reviews/market/it-security/vendor/cloudflare)

    I don’t want to play your fatuous games anymore.

    You just don’t like being corrected.

  14. Hemidactylus says

    The only hate on Cloudflare I currently have is when they tell me Pharyngula isn’t reachable. All tech companies are sus in some way, but I haven’t yet gotten all that worked up on Cloudflare. I do get a bit perturbed when the blog isn’t reachable. Glad it’s back.

    I’d like to give an oddly placed shout out to deer pose (aka 90/90s). I spent a good part of time reading the comments doing that. I’m very asymmetric for some reason deep in my hips so one side ain’t fun. Also a shout out to farmer carries. Just bought some heavier dumbbells and been exploring that and Romanian deadlifts which are a great complement to Bulgarian death squats. Got Eastern Europe covered in my routine 😀

  15. Hemidactylus says

    I’ve heard kettlebells are a thing, but kinda seems like a cult to me. I saw a video of Rogan swinging them so they could be sus.

  16. CompulsoryAccount7746, Sky Captain says

    Unrelated to the FTB outage. PHP had some nasty vulnerabilities this week. Update PHP asap.

    Kevin Beaumont (Cybersecurity journalist):

    Fixed versions will be 8.5.12, 8.4.27, 8.3.36, 8.2.35. All others are end-of-life but are still vulnerable—the vulns date back to versions from 2004 onwards. A really good mitigation for this btw—AppArmor or SELinux. […] it’s one of the biggest vulns in years.
    […]
    FWIW I think PHP team did a good job […] OpenAI ambulance chasing the PHP vulnerabilities and pretending they found and disclosed them (they did not) is disgusting.
    […]
    PHP Foundation’s budget is about 700k a year. It’s used on something like 40 million websites. Almost no company who uses it contributes anything and many people work on it for free.

    The latest trend is cybersecurity and AI vendors have realised they can vampire off its security for publicity (money).

    Perhaps the world would be a better place if those orgs who rely on it just contributed sponsorship to the foundation, and vendors didn’t monetise vulnerabilities for themselves.

  17. John Morales says

    CA7746, related to your usual context-free opinion piece extracts:

    Unrelated to the FTB outage. PHP had some nasty vulnerabilities this week. Update PHP asap.

    Note that it was found via no-good, useless AI, which always lies and outputs slop according to some.
    cf. https://www.wiz.io/blog/php2shell-php-rce

    In this post, we explain how frontier AI helped us find and fix vulnerabilities that had remained hidden in the PHP codebase for more than a decade. The issues described below all stem from key architecture decisions, some made over a decade ago, and were grouped under two CVEs, with severity reaching CVSS 9.2. The work grew into a cross-industry effort to protect customers, find variants, deploy detections and mitigations, and verify a patch in time for coordinated disclosure.

    From the horse’s mouth, that is.

    OpenAI ambulance chasing the PHP vulnerabilities and pretending they found and disclosed them (they did not) is disgusting.

    You endorse that self-serving bit of bile?

  18. CompulsoryAccount7746, Sky Captain says

    Jimmy Wylie (Malware analyst)

    Framing the php2shell story as “AI uncovered php2shell” is like hearing an awesome song and crediting the guitar instead of the player. It’s just incorrect and steals credit from a good researcher.

    The underlying bug was discovered in 2020, a related one in 2024, and they were both rediscovered again through exploitation in 2026 by a human researcher. The researcher used LLM tech to help with parts of the process, not all of it, and says so in the writeup.

    From that writeup:

    No special harness, or access to models that weren’t generally available to everyone were used in the initial bug discovery. I focused on something I had a background in, knew its vulnerability history well, and held a theory that specific bugs lurk within the codebase. This shows that “know your target” is still as applicable as ever, even in the LLM dominated VR age. No model is simply capable of unearthing “all bugs” with simple prompts.

    There are some practical research implications here. LLMs may have ingested public historical vuln data, but in their default variant analysis tends to move from the newest and most prominent bugs toward nearby older code. They do not reliably begin with obscure twenty-year-old exploitation primitives. Keeping your own mental catalog of a target’s vulnerability history lets you prompt that direction explicitly.

  19. John Morales says

    [OT]

    CA7746, you do amuse. You seem to dispute, but you do not.
    Thing is, AI is a tool. Look at how it was done.

    The researcher used LLM tech to help with parts of the process, not all of it, and says so in the writeup.

    Well, does that in any manner whatsoever contradict “In this post, we explain how frontier AI helped us find and fix vulnerabilities that had remained hidden in the PHP codebase for more than a decade.”?

    AI is a tool. The tool was used.
    The tool did not do it, no.
    But without the tool, would it have been done as soon?

    The underlying bug was discovered in 2020

    And the actual codebase primitives were announced on 19 August 2026, using AI.

  20. StevoR says

    Phew again! Again, glad to have this blog and all the other FTB ones back and really hoping they stay back this time.

Leave a Reply