5 ms·
My first thought when Musk bought Twitter was "His engineering has team figured out content moderation" but I've seen no evidence or talk of that. So I think it
by labrador 4y ago
My first thought when Musk bought Twitter was "His engineering has team figured out content moderation" but I've seen no evidence or talk of that. So I think it's true that he impulsively bought himself his worst nightmare.
- datalopers 4y agoHe's going find it's tough growing a company without a constant flow of government contracts and subsidies.
- dsfyu404ed 4y agoI'm pretty sure every MBA ever is laughing at the implicit assumptions in the article and these comments. It might not be pretty but there are other ways than growth to make a company more profitable/valuable...
- pmoriarty 4y agoI'm not sure if Musk's goal is to make Twitter profitable or to make some sort of ideological point.
- datalopers 4y agoAbsolutely. I’m talking specifically about Elon Musk. He only knows how to run gov funded organizations. Just like his buddy Thiel and Palantir.
- 93po 4y agoYou mean like he did with Zip2 at over $300 million, and X.com at over $1.5 billion? As far as I know Starlink isn't getting subsidies either. Tell us about your experience at growing companies - do you know something about this that Elon doesn't?
- datalopers 4y agoHi Jason.
- 93po 4y ago?
- qualudeheart 4y agoBiz constraints on Twitter don’t really matter to Elon. I take him at his word that be bought Twitter for ideological reasons.
- andsoitis 4y ago> bought himself his worst nightmare. His twitter profile currently says "Chief Twit". Teenage boy level humor but the joke is on the joker.
- KeyBoardG 4y agoEven if Twitter is a complete financial loss (and I think he will lose a lot of it), Elon will still be a billionaire. His personal life will feel zero of that impact.
- uni_rule 4y agoHe doesn't know what he's doing unless his main goal was always just to distract a bunch of people away from something bigger. On the concrete end I think he'll eat shit like Verizon/Oath/Yahoo did when they lost $0.8 Billion dollars on from their initial purchase price reselling Tumblr after ~6 years of ownership after running its userbase into the ground and scaring away all of its most competent original staff.
- andsoitis 4y ago> His engineering has team figured out content moderation Which of his engineering teams? Tesla? SpaceX? I don't see why they would be working on content moderation.
- toomuchtodo 4y agoHot take: there is overlap between Dojo (Tesla’s self driving ML system they feed data into to tune “FSD” models) and ingesting Tweets and signal to automate content moderation. How good will it be? ¯\_(ツ)_/¯ DALL-E is pretty good, let’s see where this goes.
- danpalmer 4y agoDojo is just machine learning optimised hardware, and it's optimised specifically for the FSD use-case which is lots of image processing. I can't imagine it having any benefit whatsoever for Tweet moderation above using some GPUs for ML. Also Musk likes to pitch integration between his companies but the reality is that there's very little.
- SilverBirch 4y agoAnd if Tesla engineers are working on content moderation for Twitter I think a few Tesla share holders would be interested in why they're bankrolling Musk's other enterprises.
- cma 4y agoSpace X bailed out Solar City with NASA money by buying its junk bonds. Tesla bailed them out with a fake shingle presentation to drum up an acquisition. In the past the board of SpaceX got angry at its workers and resources working for Boring Company: > The arrangement alarmed some longtime investors in SpaceX, including its largest outside backer, Peter Thiel’s Founders Fund, some of the people said. The investors learned in recent months that despite the diversion of SpaceX resources and staffing to the fledgling Boring startup, it was Musk who was in line to receive almost all of any future profits, these people said. https://www.marketwatch.com/story/new-questions-over-elon-musks-use-of-spacex-resources-for-boring-co-tunnels-2018-12-17 https://www.marketwatch.com/story/new-questions-over-elon-mu... It was basically embezzlement, but they ended up with a quiet settlement giving SpaceX some shares: > In early 2018, The Boring Company was spun out from SpaceX and into a separate corporate entity.[27] Somewhat less than 10% of equity was given to early employees, and over 90% to Elon Musk. Subsequent concerns by SpaceX shareholders resulted in a December 2018 reallocation of 6% of The Boring Company's equity to SpaceX. https://en.wikipedia.org/wiki/The_Boring_Company#History https://en.wikipedia.org/wiki/The_Boring_Company#History
- pjkundert 4y agoPerhaps Elon has heard of K-Means Clustering, and you haven't... "Solving" content moderation is a losing battle -- there is no way to humanly moderate anything at the rate it can be produced, and Elon knows this. It's insane, so he probably won't do it. But, allowing content to be produced, but making sure no real person ever sees it, until the content producer earns their way into "your" group -- now that can be automated. Elon has access to Exascale hardware, so K-means Clusters with arbitrarily large dimensionality is at his disposal. Ya figure he might not have thought this through?
- labrador 4y agoOh sure, we can talk about Voronoi cells all day, but the Weber problem indicates an increasing number of k-medians and and a decreasing number of k-medoids. Anyone with the slightest familiarity with the Rocchio algorithm knows that.
- pjkundert 4y agoSo, if you know that Fermat-Weber had fast approximation solutions (O(logn)) published by 2003 (Bose, Mahashwari & Morin), and Rocchio "training" is linear O(|D| + |C|) (where D is the Document set (tweets) and C is the Class set (agents)), why wouldn't Elon maybe be able to use his existing automatic "Labelling" AI workflow to implement something like Rocchio and/or Fermat-Weber? I didn't meant to be rude; but, you claiming "I've seen no evidence of that" also seemed rude. Sorry for responding "in kind", and not assuming the best. But, seriously; I immediately thought "Wow, I've seen evidence of that", straight from Elon's latest "AI Day" presentation... My theory on "why" so many smart Google / Meta / Twitter / ... people have not implemented K-Means Clustering? Because it gives power back to the Agents in the system -- they can't shove their pre-selected content in your face to drive Ad clicks. I think Elon doesn't care so much about that. Or, sees how giving back authority to the agents will eventually attract orders of magnitude more agents / interaction into the system, offsetting the per-agent reduction in Ad revenue. EDIT: Sorry, O(nlogn) deterministic or O(n) time-random approximations for Fermat-Weber.
- mikkergp 4y ago> His engineering has team figured out content moderation Depending on how you define content moderation this is literally not possible, 1. Because content moderation is adversarial and has to adapt and 2. Content moderation is subjective, as the original article says content moderation is the product that companies sell, and it will likely have to change as time goes on (I’m not just talking “what’s popular/offensive” but like do you allow videos, how long are the videos, do videos get prioritized above text or audio, 280 character vs 144 character tweets, editing tweets, dislike button, etc.