It would be helpful if you answer the question about web api usage, most of that is not relevant.
The only suggestion I see there from a quick skim that would avoid the above is for customers to set up a google maps proxy server for every usage with adds security and hides the key. That is completely impractical suggestion for the majority of users of embedded google maps.
They are talking about pnpm (which they said would be the uv equivalent for node, though I disagree given that what pnpm brings on top of npm is way less than the difference between uv and the status quo in Python).
Prices are not public, but comments over the years have hinted it’s in the $500K to $1mm range.
Their hardware is multiple generations behind at this point, however. I wonder if they’re starting to reduce the price because it’s hard to justify paying so much for old hardware. They could just be targeting customers who don’t care as much about performance or efficiency as they do the software stack.
Being a few generations behind is kinda par for the course for any server hardware that's put in production, this is not a gaming PC build. Hopefully they're working on bringing their hardware up to date, since efficiency is a key consideration for the class of workloads they're aiming at.
> Being a few generations behind is kinda par for the course for any server hardware that's put in production
No it’s not. Normally if you’re buying server hardware you don’t start with a CPU that’s already 5 years old and last-generation RAM that isn’t even manufactured at scale any more.
CPUs have advanced a lot in recent years. The jump from Zen 3 to Zen 5 is very substantial.
From what they say and their podcast its pretty clear that they already have the next generation sled and likely already sell it to existing costumers, its just not on the website.
I just talk about some comments I read some time ago, so take this with a grain of salt. It was my understanding that if you were spending about $300k-$500k a year in cloud services, it would make sense this type of solution, so the expected price would be something between $500k-$1M depending on the configuration
I am getting the AI Agents to build an expense tracker Telegram. I would like to have one myself and among my family members since we are heavy Telegram users. I am also using this as a way to learn more about the AI Agents (what they are good at, their limitations, etc) with (hopefully) proper guardrails, guidelines, checks, etc.
As you may see from the git history and "contributors", it's mostly Claude and AMP making the changes.
I am not entirely sold on these agents and not particularly excited by these. But I also feel that I can't afford to sit out this transition so here I am...
