• 1 Post
  • 34 Comments
Joined 1 year ago
cake
Cake day: June 14th, 2023

help-circle
  • Asking the person you’re debating to look up your own citations is certainly one way to converse. But ok, let’s go for it.

    In Aug 2023, Forbes published an article describing the proposal of “unfettered access” you referred to:

    https://www.forbes.com/sites/emilybaker-white/2023/08/21/draft-tiktok-cfius-agreement/

    In June 2024, the Washington Post reported that the Committee on Foreign Investment in the United States (CFIUS) turned down the proposal and includes some broad reporting as to why:

    https://www.msn.com/en-us/news/politics/tiktok-offered-an-extraordinary-deal-the-u-s-government-took-a-pass/ar-BB1nfAcE

    The article isn’t very technical, but it mentions some interesting responsibility angles that the US wouldn’t want to back themselves into:

    • throwing open some, but not all, doors to server operations and source code creates a mountain of work for the government to inspect, which would be a workload nightmare
    • the US government’s deepest concerns seem to be about what data is going out (usage insights on the virtuous side and clipboard/mic/camera monitoring on the ultra shady side) and data coming in (bespoke content intended to influence US residents of China-aligned goals). Usage insights are relatively benign from national security perspective (especially when you can just mandate that people in important roles aren’t permitted to use it). Shady monitoring should be discoverable through app source code monitoring, which you can put the app platforms (Apple, Google, whoever else) on the hook for if they continue to insist on having walled app gardens (and if you trust them at all). The content shaping is harder to put your finger on though, since it’s super easy to abstract logic as far out as you need to avoid detection. “Here, look at these 50M lines of code that run stateside, and yeah, there are some API calls to stuff outside the sandbox. Is that such a big deal?” Spoiler: it is a big deal.
    • the US can’t hold Byte Dance accountable so long as it remains in China. Let’s say the US agreed to all this, spent all the effort to uncover some hidden shady activity that they don’t like (after an untold amount of time has passed). What then? They can’t legally go after Byte Dance’s foreign entity. The US can prosecute the US employees, but it’s totally possible to organize in such a way that leaves those domestic employees free from misdeeds, leaving prosecutors unable to enforce misdeeds fairly. It’d be a mess.

    The second article explains this somewhat, but I’m admittedly painting some conjecture on top regarding how a malicious actor could behave. I’ve got no evidence that Byte Dance is actually doing any of that.

    But going back to the “influence the public” angle, I’m struggling to see how different TikTok is versus NHK America (Japan’s American broadcasts) or RT (American media from the Russian standpoint) aside from being wildly more successful and popular. But I guess that’s all there is to it.

    I’d prefer our leaders also be transparent with us regarding their concerns about TikTok. The reductive “because China!!1!” argument is not compelling on its own.






  • 12 year SDE + 12 year TPM vet here.

    Do everything you can to help your software engineers (or whoever is doing the work) have as much focus time as they need. Buffer your meetings and questions to one chunk of time per day. Encourage them to block-out and protect their focus time. And encourage the team to keep office hours so they can still make themselves available to others, but in a controlled way.

    Be transparent with the business’s goals and frustrations you are facing. There’s an attitude (often among inexperienced devs) that PMs are good for nothing; just an interface to the rest of the business, and a source of where tasks come from. And some certainly are that, but a good PM is worth their weight in gold.

    Find a good mentor, and start thinking about your next career step now.











  • Read the article.

    Machine learning and interpretative output are tools; just like the automobile, the spreadsheet and photoshop.

    The introduction of new tools means there will be fewer people manually doing the things that machines can do more efficiently. The introduction of digital spreadsheets decimated the market for paper bookkeepers, but the need for accountants (people who could utilize the new tools) exploded.

    I don’t know enough about modern animation production to speak authoritatively about this, but I’m imagining Katzenberg is talking about jobs like inbetweeners and other kinds of admittedly skilled labor that can be lazily farted out by machines. No QA for lazy productions, QA and varying levels of tweaks for high production value work, and all-by-hand for only the most rare auteur works. And most animated works are in that “lazy production” category. It’s gonna look like shit, everyone who cares will notice, but most of the people buying won’t care.

    What this also means is that money will stop flowing to high-manual-effort works. The real creative, ground breaking stuff is going to come from either people utilizing the new tools in new ways, or old established artists who refuse to change (Miyazaki, Bill Plympton, Yuri Norstein & Francheska Yarbusova, etc).