Could you explain how supporting multiple languages increases the parameter count so much? I'm genuinely curious.
LLMs seem to be comfortable with hundreds of programming languages, DSLs and application specific syntaxes so how does supporting a couple more natural languages become so expensive?
I see how more training data would be needed, but I don't understand how that maps to a greater parameter count.
you are underestimating the impact of programs like this.
The CEO will quickly spot broken processes and inefficiencies. Ass kissing can't make up for business problems. Beyond that, by working in a different store, gets a feel on how generic the approach can be, and where they need local optimizations.
The only think you are taking out of the consideration is toxic lower management and asshole colleagues, but that's not something I'd expect the CEO to fix directly himself. If anything he is just bored of the ass kissers and good news shows, and wants to regain a feel of reality
> The CEO will quickly spot broken processes and inefficiencies.
They will quickly spot what they perceive to be broken processes and inefficiencies, but they don't have real experience of the actual work at that level to be able to safely move Chesterton's fence.
Nope. The role of executives at large companies is to most effectively move resources away from labor and towards the owners of capital. They decide where to open the next branch to crush competitors who treat their workers and customers better; they negotiate deals with suppliers so those suppliers and their workers have fewer resources; they coordinate with other large competitors to hamper workers' attempts at improving labor conditions.
This, to me, doesn't qualify as work. Large company executives create no value. They only take and store value from those who actually perform real work.
you're conflating work and, uh, "activity", so you're just talking past the OP instead of understanding them.
a common usage definition of work being used here is in terms of creating "use value". the ceo's activities do not. they do generate profit for ownership though. which is not the same thing as "use value" and serves a different purpose in society.
Literally every time I've moved up the chain in my career I have done less work. I have seen the same in peers who have surpassed me. I have no reason to believe this changes by the time you reach CEO level.
The bit where they bring in a pro CEO in Silicon Valley who spends most of his time messing around with his horse-breeding hobby instead of running the business is basically just a documentary. Like most of the rest of the show.
This is how CEOs often manage to "do so much". All kinds of charity involvement, "advisor" work on other businesses, sit on boards, et c, then blog about how they still find time for their family and/or staying fit despite being so "busy". It's not because they're super-humans working 100 hours a week with perfect time-management discipline, but because a lot of them don't really do jack-shit for any of their "jobs".
the execs definitely miss out on a lot of money because bonus programs are typically multi year.
There are always
-short term (how did we do this year, the classic everyone gets)
-long term (bonus defined today, paid out over a few years if you meet the strategic goals, typically director and exec level, we are easily talking 100k+/year here in bonus money)
bonusses. So, yeah, still a financial letdown for them. But they still got paid quite a lot of money of course. The average swiss person had no stakes or no culpability at all (except maybe choosing them, clients could have walkd away)
But they've already made out like bandits in the meantime, and can move on to the next company to raid (I mean, be an executive of).
Like you can make $5 million a year now, but have things suddenly end on you at some point when things come crashing down, but probably not for several years, or you can make $1 million a year, but play it safe, and in five years get a $7 million long-term bonus.
Might as well get as much out of the company now and trust the government will bail you out (or not, what do they care, they can just move on to the next thing, and even if not, they've already made a shit ton of money, they could just retire, no big deal).
In my country it’s not illegal when you punch somebody in the face, but we have another deterrent that works very well. If you punch somebody in the face, then he pulls out his wallet and gives you $50, then presents his face for you to punch again if you’d like. Once you stop punching, he gives you $100. So you see, assailants quickly learn that face punching is an action of grave consequence and loss, and the fact that everybody here is punching each other in the face all the time is probably because of something else.
They are brilliant at marketing (look at DALL-E). But then stable diffusion comes along etc and they need to prove their worth against competition. I am afraid Microsoft is not handling this well
a 400 month customer is really a tiny fish, don't expect special treatment without support packages. Do you even have an enterprise contract with them? The bronze contract can be had from 0 spend, and gives the option to pay through invoicing and have support. Or are you just going through credit card which is intended for hobby use?
I am not defending them, but this is totally on you.
What was the missing information? Lack of identity verification or credit card verification? Again that's on you. You should have also received a warning in the cloud console and if you hvae the cloud app on your phone, there as well.
> I am not defending them, but this is totally on you.
You are defending them. Unless OP is disingenuous, the reputational loss of an unhappy customer is hard to shake. Customers will make mistakes, if the cloud provider doesn't have automated systems to catch and prevent those mistakes then they aren't doing well in customer service.
A similar thing happened to me with my personal Azure spend, but I felt they handled it much better simply by communicating early and often when there was a problem processing payment. I had time to respond and fix my subscription, so I stayed a happy customer.
Fortunately traditional hosting providers like Hetzner won't suspend me even if I pay $10/month on average. That's why I avoid Google Cloud Platform, Microsoft Azure and AWS like a plague. I prefer to have a freedom to switch my hosting provider at any time, even if that sometimes means more infrastructure work. That kind of attitude to customers, as GCP usually has, is just not for me, thanks.
I don't understand you. I have said that Hetzner is not going to suspend my account (or deprive me of customer support) on the grounds of paying too little per month. What kind of data point you need? I am the data point. I'm their customer since 2015.
The general point I was trying to make is not an advertisement of particular hosting provider, but the fact that traditional hosting providers are totally fungible. If tomorrow this company goes bonkers and starts doing stupid things, then I just migrate to OVH, DigitalOcean, Vultr or any other competitor on this market. But unlike traditional hosting providers, there is no market among big cloud providers, as they are not fungible on the customer API level, so each of them is actually a monopoly. And monopoly is generally no good for customers. The rest can be deduced from this principle.
I wouldn't say that... there are a LOT of sites that operate at under $1k/month spend. If you lose 10k customers for one big fish, is it worth it? I know it's not apples to apples, but still can really matter. And just maybe it's not worth supporting a bigger player in that case altogether.
And for the bigger fish, if it's me, I'm one to push for complete containerization and orchestration beyond simple DBAS service for PostgreSQL, depending on needs for scale. Which absolutely means switching can be easy enough.
The only advantage the big cloud providers have is the services for the lower-middle tier customers.
I've just coded my latest site using aws but I'm going to be teaching myself terraform in the coming week so I can rewrite the infrastructure as code so I can move providers easily if I ever have any problems in the future.
Terraform is not going to make you cloud-agnostic. If you switch cloud providers, you'll still use Terraform but you will have to rewrite much, if not all, of your TF code.
And even if I don't, I imagine rewriting a terraform workflow is probably easier than taking cloudformation code and then rewriting it into whatever the equivalent is in google or microsoft land. Atm I'm pretty happy with aws and don't see myself moving but I like to have the option if needed.
If you can be banned from the platform at any time just for having the audacity of using your credit card, maybe they should have a big yellow warning box when signing up with a credit card
it's simple: Many companies are on Azure for some reason. (mainly because of preexisting microsoft contracts, and because its so tightly coupled with office that if you want office at the enterprise level, you always need a little of Azure)
Now, those companies need data infrastructure. Azure ML studio is horrible. Azure data factory is somewhat okay'ish, but far from ideal. Azure Synapse is a steaming pile of junk that doesn't seem to improve. Azure data lake is kinda functional as a data lake, but lacks any decent ways to analyze that data.
Spinning up jupyter notebooks is not something most IT teams understand, they want a managed service.
As far as I know, Databricks is then the only solution offered by Microsoft. That's why they get such a high adoption rate. Not because of merit, but because they are the only somewhat acceptable solution on Azure. Good enough to just work. Not good enough to gain a competitive data advantage.
I hate what the world has come to, but sadly IT procurement has too much influence and CTO's dont understand the modern world enough. They just care about the big Microsoft contract, because it makes them look good (look at all these savings!)
If Microsoft would showcase Teradata or Dataiku or a bit of Snowflake, all of those run through the azure marketplace anyway,... we would be better off. But consultants and premium partners naturally only care about Databricks
Well said. It is a sad state of affairs.
For those of us who can, push back against IT procurement and show leadership the value (and likely cost reduction) when using tools outside the norm.
And yes, the Azure Marketplace has several great solutions for data processing.
