RE: https://hachyderm.io/@thomasfuchs/116440570565467277
"These long-running, parallelized workflows can yield great value, but they have also challenged our infrastructure and pricing structure: it’s now common for a handful of requests to incur costs that exceed the plan price! These are our problems to solve. The actions we are taking today enable us to provide the best possible experience for existing users while we develop a more sustainable solution." — Joe Binder, VP of Product
If people are already _commonly_ exceeding set monthly quotas/costs with just a handful of requests, you all have got much bigger problems, customers/users too (i.e. becoming deeply dependent on externally controlled & rented infrastructure from day one of your project/startup, plus slowly damaging your own cognitive abilities and a fast growing maintenance nightmare)! And this all is just the beginning... 🍿
Keeping up with other news about growing mineral, energy & supply chain shortages, the illusion that this might just be very short term, and the resulting cost bumps needed to produce and provision all the outstanding data center capacities at the heart of the above decisions (and similar ones made by Anthropic just days ago), I really don't understand where this optimism of AI boosters/users is coming from or on what planet they're living on...
genAI/agenticAI is a shoddily implemented/managed/secured technology & mentality for a computing-illiterate consumerist society, for an age of infinite resource abundance with stable climate/politics, all of which we're leaving (or have never really been in)! And that departure is partially driven/accelerated by said tech...
Instead of funneling trillions into this particular tech, we need focus/investment into methods, infrastructure, research/education, relationships, and above all, a culture focused on how to make computing magnitudes more sustainable, accessible, independent and which allows us to do generally more with less. Or, at least not with magnitudes more, just for some dubious, unspecified (and meaningless) measure of improved "productivity gains"... With almost no one ever questioning what this supposed extra productivity is even good for...






