So this was one of those weeks where I kept refreshing the news and just kind of sighing. Half my Twitter (X, whatever, I still call it Twitter and I'm not stopping) was airport departure boards full of red text on Friday, and by Monday everyone had moved on to arguing about something else entirely. I wasn't personally stuck anywhere, thankfully, but a friend of mine spent nineteen hours in Newark because of it, sleeping against her rolling suitcase like a lot of other people that weekend. I'm not going to rehash the whole outage thing here because you've read six other posts about it already this week, probably from people who understand the technical side of it better than I do anyway.
What actually caught my attention, buried under all that, was Google quietly admitting on Monday that it's giving up on killing third-party cookies in Chrome. Again. This is like the fourth time.
If you don't follow this particular corner of ad-tech misery: Google announced back in January 2020 that Chrome would phase out third-party cookies, the little tracking crumbs that let ad networks follow you from site to site, by early 2022. Then it got pushed to 2023. Then late 2024. Now, as of this week, they're basically saying never mind, we'll just add a new prompt where users can choose, and Chrome keeps the cookies around indefinitely. Four years of "Privacy Sandbox" blog posts, API proposals, developer origin trials, congressional testimony practically, and it lands on: eh, we'll ask you nicely instead.
I called this back in 2021 in a comment on some now-dead forum, so I'm allowed a small victory lap. Nobody who actually makes money from ad targeting was ever going to let Google turn off the thing that makes ad targeting work, not when Google itself runs the largest ad business on the planet. The whole Privacy Sandbox thing always smelled like a company trying to look like it cared about privacy while making sure nothing that mattered to its bottom line actually broke. Regulators in the UK (the CMA) were breathing down their neck about exactly this, worried that killing third-party cookies would just hand Google's own ad stack even more power since everyone else would have to scramble to adapt while Google's first-party data doesn't need cookies in the first place. Kind of funny that the antitrust worry ended up being the thing that let cookies survive.
I run uBlock Origin and have third-party cookies blocked at the browser level anyway, have for years, so functionally none of this changes anything for me personally. But it's a good reminder of how much of "we're fixing the internet's privacy problem" from big platforms is really just "we're fixing it in a way that's convenient for us, on a schedule we control, and we reserve the right to not fix it at all if it stops being convenient." Apple did the tracking prevention thing in Safari years ago and just... did it. No four-year rolling deadline, no origin trials, no "Privacy Sandbox" branding exercise. It wasn't perfect either, but it happened.
Also worth a mention, since it got a bit drowned out this week too: Meta put out Llama 3.1 on Tuesday, including a 405 billion parameter version, and is calling it the biggest open-weight model release so far. I haven't had a chance to actually run it (405B is not exactly a laptop-friendly download, my home server would just cry) but I've got it queued up to poke at once I find a rental GPU box that won't cost me a fortune per hour. If anyone's got a cheap Runpod or Lambda setup they like for this kind of thing, genuinely, email me, my inbox is open and I'm tired of guessing between providers.
Anyway. Cookies aren't dying. The web's tracking economy gets to keep limping along a while longer, propped up by a company that also happens to be the biggest beneficiary of it staying alive. Shocking absolutely nobody who's been paying attention since 2020.