4 ms·
If this is Linda Yaccarino's first major policy decision as CEO then it's not looking good. I really don't know what they (corporate) can do at this point.
by andyjohnson0 3y ago
If this is Linda Yaccarino's first major policy decision as CEO then it's not looking good. I really don't know what they (corporate) can do at this point.
- threeseed 3y agoMusk is in charge of Product and Technology so this is his decision. She is much more akin to a CEO at Large running around closing ad deals.
- meepmorp 3y agoImagine trying to do your job closing ad deals and the owner decides to block public access without prior notice. Nothing says "we're a trustworthy partner who won't damage your brand," like major changes made seemingly at random.
- kogepathic 3y ago> Nothing says "we're a trustworthy partner who won't damage your brand," like major changes made seemingly at random. This has been Twitter's Modus operandi from well before Musk acquired them. They announced new features through Tweets (remember Fleets?) with glaring issues at launch (no authentication required to view a Fleet) and without any corresponding API documentation. The GraphQL that backs their website has more features and bug fixes than their paid API! They also famously do not have a staging environment, which frequently caused issues when new features were deployed or tested with a subset of users. Twitter has never been a trustworthy or reliable partner. They are just more of an obvious dumpster fire since the Musk acquisition.
- fragmede 3y agoStaging environments are really hard at scale and I don't think anyone's really cracked that nut yet. The coordination between umpteen teams means that if there even is a staging, the test data is not totally representative, the systems are wildly out of sync, and it's either too stable, or too unstable. So you deploy to staging, which doesn't really give any signal other than it doesn't (or maybe it does) crash the kernel, so you release to prod with a gradual rollout and just pray. Twitter's got its share of issues, but not having staging isn't one of them because there's not much value for massive distributed systems like that to have a staging environment, not relative to the effort required, anyway.