413 karma · joined September 9, 2010
Because from my perspective that's where microservices can especially break down: attrition or layoffs resulting in service ownership needing to be consolidated between fewer teams, which now spend an unforeseen amount of their time on per-service maintenance overhead. (For example, updating your runtime across all services becomes a massive chore, one that is doable when each team owns a certain number of services, but a morale-killer as soon as some threshold is crossed.)
This part is pretty hype-y. Old-fashioned deterministic web app vulnerability scanners can of course be used to make multiple requests per second. The limiting factor is probably going to be rate-limiting on the victim's side / # of IP ranges the attacker can cycle through, which would apply to the AI-driven vulnerability scan too.
I think most of the ones from previous years are all in people's junk drawers, couches, etc., and only go back into circulation when someone decides to dump them into a Coinstar machine. Retailers are already reporting shortages.
(And it seems like there's a physiological basis to both cases, it's just that one involves endogenous chemicals and the other doesn't)
People will complain about paywalls and they will complain about ads. The money needs to come from somewhere. For ad-supported sites, increased ad blocker use means higher saturation of ads for everyone else.
It's always very nice to see a nonprofit journalism outlet with no ads, but donor funding does not scale up to a large newsroom.
I do not have at all the right background to evaluate this research so treat this opinion for what it's worth, but it seems incautious for the authors to close with this note near the end. People like RFK are looking for an explanation for that 20-fold increase. But the hazard ratios in the studies with positive results seem to be along the lines of 1.05-1.20. They do also note changes in diagnosis criteria before this sentence, but it still seems like if they're going to mention a 20-fold increase, they should be even more explicit that any association with increased Tylenol use could only ever explain a very small part of that.
But at end of day it's up to the site operator, and any server or reverse proxy provides an easy way to block well-behaved bots that use a consistent user-agent.
The Big Tech bots provide proven value to most sites. They have also through the years proven themselves to respect robots.txt, including crawl speed directives.
If you manage a site with millions of pages, and over the course of a couple years you see tens of new crawlers start to request at the same volume as Google, and some of them crawl at a rate high enough (and without any ramp-up period) to degrade services and wake up your on-call engineers, and you can't identify a benefit to you from the crawlers, what are you going to do? Are you going to pay a lot more to stop scaling down your cluster during off-peak traffic, or are you going to start blocking bots?
Cloudflare happens to be the largest provider of anti-DDoS and bot protection services, but if it wasn't them, it'd be someone else. I miss the open web, but I understand why site operators don't want to waste bandwidth and compute on high-volume bots that do not present a good value proposition to them.
Yes this does make it much harder for non-incumbents, and I don't know what to do about that.
The fact that LLMs can generate this kind of output is not incompatible with the OP's argument about how useful they can be when combined with agents and code review. But search is the lowest-friction way for anyone to interact with LLMs currently, and right now it's a pretty poor advertisement.
I think this is almost already the case now. Services like Cloudflare do a pretty good job of classifying primitive bots and if site operators want to block all (or at least vast majority), they can. The only reliable way through is a real browser. (Which does put a floor on resource needs for scraping)