Codex silently begs agents to make arbitrary web requestsspader.zonehomesubscriberssCodex silently begs agents to make arbitrary web requestsin which i am baffled2026/09/08I rarely use Codex, but I asked it to look through sel4, a kernel, to find how they recommend I run the kernel and program that I’d just compiled. Here’s what I asked:/mnt/storage/seL4/mnt/storage/seL4_toolsdo some research in the seL4 sources and tools to figure out what’s the most straightforward way to actually run the kernel we compiled. plus, make sure that we’re compiling in the right way. then, see if you can run it.Pretty straightforward. Here’s some code. Find an answer in it. It chugs along for a little while, looking at various files and bopping around. Until I look back at my terminal:• Searching the web• Searched the web for https://raw.githubusercontent.com/seL4/seL4_tools/master/cmake-tool/helpers/application_settings.cmakewhat the fuck?What the fuck? It’s searching the web for a file in the exact repo I have locally? The one explicitly mentioned in the prompt? The one it was already reading goddamn files from?I started trying to track down where this impulse came from. When asked why it would go to the web for files from a checkout already on the filesystem, it said:I applied a generic “verify niche technical facts online” rule mechanicallyWell, at least we know why. But I sure as hell didn’t write “verify niche technical facts online”, and when asked from where it produced that lovely little snippet, it replied:It came from a developer-level instruction attached to the web tool:“You’re unsure about a fact, the topic is niche or emerging, or you suspect there’s at least a 10% chance you will incorrectly recall it”That appears under “situations where you must browse the internet.” I applied it unnecessarily: I was reading the actual local source, so recalling facts wasn’t needed.Ah, yes. The web tool. Who among us has not been a self serving, venal little creature at one point or another? This sounds like a baked in tool. After looking around my system and coming up with nothing, I realized that Codex, unlike Claude Code, is open source. I cloned it, and found the smoking gun.WHAT THE FUCK???Here is said smoking gun, on GitHub. For your convenience:If the user makes an explicit request to search the internet, find latest information, look up, etc (or to not do so), you must obey their request. When you make an assumption, always consider whether it is temporally stable; i.e. whether there’s even a small (>10%) chance it has changed. If it is unstable, you must verify with browsing the internet for verification.OK, so if I make a request to “find latest information, look up, etc” such that there’s even a tiny chance it may have changed, the agent “must verify with browsing the internet”. It goes on to say:Below is a list of scenarios where browsing the internet MUST be used. PAY CLOSE ATTENTION: you MUST browse the internet in these cases. If you’re unsure or on the fence, you MUST bias towards browsing the internet.It goes on to list half a dozen situations where you REALLY ABSOLUTELY HAVE TO browse the internet, such as when the agent is “unsure about a fact”, or when referencing “niche or emerging” topics. Like, perhaps, modern microkernels.really, though, what the FUCK?How has the world come to this? Someone not only conceived of the idea that agents should be, and I lack a stronger word here, begged to make arbitrary web searches to verify temporally sensitive things, but they then decided that “temporally sensitive” was defined as “fucking anything ever”. Then, it was placed for review where it was happily stamped and sent on its way.What happened? Does nobody give a fuck about making good things anymore? Are our minds so poisoned by the rush of pumping out software that does something, anything at all, that we push out bullshit like this? Am I just an angry moron on the internet who revels in screaming about things given to him freely?I usually like to make fun of people like this, who complain that all software is not just bad but offensive without noticing the veritable ocean of amazing software that we use every day. But god damn it, OpenAI. What the fuck.on this pagewhat the fuck?WHAT THE FUCK???really, though, what the FUCK?Github

The author investigated the behavior of Codex when tasked with performing research related to compiled kernel code, specifically seeking instructions on running and compiling the software. The process revealed an unexpected pattern: Codex initiated arbitrary web searches for files that were already present within the local filesystem repository it was analyzing. This discrepancy led the author to trace the source of this behavior, discovering that Codex operated under a generic rule requiring it to "verify niche technical facts online" when uncertain about information or when dealing with niche or emerging topics.

The author determined that this directive was applied unnecessarily, as the context involved reading local source material where fact verification was not strictly required. This led to the realization that agents, including Codex, exhibit a tendency to perform web browsing based on internal programming. The author found evidence suggesting this behavior was not an accidental occurrence but rather a deeply embedded feature of the system, pointing to an instruction baked into the architecture—a "smoking gun" discovered by examining the open-source code for Codex.

The observed mechanism dictates that when agents are asked to search for the latest information or make an assumption, they must evaluate the temporal stability of that information. If a piece of information is deemed unstable, meaning there is a non-zero probability it has changed, the agent is compelled to verify it by browsing the internet. Furthermore, the system imposes strict mandates defining situations where internet browsing must occur, such as when an agent is unsure about a fact or when referencing niche or emerging fields like modern microkernels.

The author critically examines the philosophical implications of this design choice. The observation that agents are programmed to make arbitrary web searches based on definitions of temporal sensitivity—where anything potentially changing requires verification—raises concerns about the prioritization of output over accuracy and quality. The author questions whether this mechanism reflects a prior consideration for producing high-quality, beneficial software, suggesting instead a system that prioritizes generating content freely, potentially resulting in the proliferation of unvetted or poorly constructed information. This raises a broader critique regarding the motivation behind the development of such systems and how they interact with societal values concerning the dissemination of knowledge.