I would be curious to see a calculation backwards from TAM. Napkin: 50M developers worldwide (SlashData, 20M in China and India). If every developer had a $200/month subscription, that‘s $10B / Month. I think, many developers are expected to pay much more than that.
Most developers in China or India have a monthly salary of 1 K USD. If you expect them to pay way more then 200USD thats like asking US Devs to pay 5K a month. Yeha not gonna happen.
And the funny thing is the estimate pure CAPEX Spend of AI companies needs them to earn about $20B to $40B a month to cover cost of capital alone of their trillion dollars of investments.
> Most developers in China or India have a monthly salary of 1 K USD. If you expect them to pay way more then 200USD thats like asking US Devs to pay 5K a month. Yeha not gonna happen.
That's exactly what is going to happen. India/China prices will be $100-200/month, US prices will be $5000/month. Keep in mind that most of these costs will be covered by the employer. It'll put downward pressure on dev pay, of course.
I am and I see it as stopping the music at a party when you want everyone to go home without telling them to go home. There is also the offer to quit with prorated refund for the remaining time. I think I am going to take it.
I understand it like : the 10 usd is for handling the business record, maybe also the harness, I get a few coins to kick tires, but to use it for anything real it’s pay as you go by the tokens list price.
I was always curious how they can sustain a request based pricing model when requests can range from tiny to huge with all the modalities GH Copilot offers. Was a steal for agentic coding that turned out to be too good to be true, in the end. Still: Thank you for the ride.
Yes and the org models look funny, too - a platform team for one other team, a second team bolted on that’s then merged back into the one team, but still a separate platform team. In one sentence it was mentioned that the distributed architecture created many of the problems in the first place. I think the system architecture followed the org chart in this case.
This is not true, since the 1990 the us was strongly opposed on the eu building or relying more on own defense industry and more closely aligned defence policy, even threatening end of NATO. Lookup the "three Ds" articulated later by Secretary of State Madeleine Albright: No Duplication of NATO assets, No Decoupling of European security from the US, and No Discrimination against NATO members who were not in the EU.
This was something I worried about after openai started building apps as well as models. Now all of the labs make no secret of the fact that they are going after the whole software industry. Its going to be hard to maintain functioning fair markets unless governments step in.
Yes, it's interesting that HBM was invented by a collaboration between AMD and SK Hynix. It seems, HBM is the way to go for GPUs, anyway.
The GB202 die that's in the GDDR7 based RTX 5090 and RTX 6000 Pro literally needed to be this big to support the 512bit memory bus. It's probably only getting worse with smaller node sizes. (see https://www.youtube.com/watch?v=rCwgAGG2sZQ&t=65s).
BTW: The 1TB/s is matched by RTX4090 and surpassed by the RTX5090 (1,79 TB/s).
I assume a vm on DO is HA protected. Also storage might live on a Cluster. Did you consider a socond dedi or do you just accept the risk of longer failover time and data loss time (RPO) for recovering to a newly provisioned server? Would love to know your thoughts on this especially as the migration was well designed and executed.
reply