dinvlad

This is the best take

This is such a deep rabbit hole! Other alternatives include CDN and residential proxies, no VPN required

Jeez. Thankfully I carefully avoided almost any and all of that political driven craziness there, as my bubble is just other solopreneurs

This. I’m on X myself a lot (because everyone else in my circles is, not because I wanted) and it’s just such a bubble. Sometimes I want to just quit it all and touch grass

Not a single VM possible to request on Azure us-east for over a month now though :-(

Right now, we can’t request even a single (1) non-beefy non-GPU VM in us-east on Azure. That’s been going on for over a month now, and that’s after being a customer for 2 years :(

The cloud is also not fulfilling its end of the promise anymore - capacity on-demand. We’ve been struggling for over a month to get a single (1) non-beefy non-GPU instance on Azure, since they’ve been having just insane capacity issues, where even paying for “provisioned” capacity doesn’t make it available.

And now, on cloud it’s the same but much more expensive and worse performance. We’ve been struggling for over a month to get a single (1) non-beefy non-GPU VM allocated on Azure, since they’ve been having insane capacity issues, to the extent that even “provisioned” capacity cannot be fulfilled ;-(

The whole business model even for OAI/Anthropic is unsustainable.. they are already running it at a huge loss atm, and will do for the foreseeable future. The economics simply doesn't work, unfortunately or not

Another one bites the dust :-( I hope at least Windsurf stays the same..

Andras (author of Serverless Horrors) knows what he’s talking about.

The amount of brainwashing that big cloud providers have done, is insane.

Saw this too, but seems a little too good to be true tbh. I don't doubt it helped them, but it's just so hard to make these agents work reliably. And the more "rules" one accumulates, the less reliably they follow all of the rules and details. Heck, Claude can't even follow a single rule of not inserting comments everywhere.. Plus, code often ends up in a mess, without strict human oversight.

Perhaps I'm overlooking something though, it would be great to see more concrete details of their implementation, rather than this high-level inspirational description..

when the "right" format and "right" time are essentially, and maybe even necessarily, undefined, then aren't you still reaching for a "magic" solution?

Exactly the problem with all "knowing how to use AI correctly" advice out there rn. Shamans with drums, at the end of the day :-)

already research allowing LLMs to use millions of different tools

Hmm first time hearing about this, could you share any examples please?

I feel like ppl just keep inventing concepts for the same old things, which come down to dancing with the drums around the fire and screaming shamanic incantations :-)

Thanks a lot for clarifying in plain language! Makes sense re basically anything paid is NOT used for training, free - does.

Off-topic, but I wish this kind of plain language doc existed for Google One vs Google Workspace as well.

Egress is $25/TB unlike their other stuff. No thanks

Surprised no one else mentioned "it's always DNS" yet :-)

Except that AWS, CloudFlare and a bunch others are also down :-O

#HugOps

So wholesome :-)

Just like Tesla "FSD" :-)

Already do :-)

Hetzner does have a cloud VPS option, which is still very affordable: https://www.hetzner.com/cloud

This is pretty heads-to-heads with EC2, in terms of how it works behind the scenes.

Yeah, they're all pretty fantastic. Can even upgrade existing servers with more RAM/disk, or store ~250 TB of data for $400/mo etc.

Not to mention unlimited free 1 Gbit egress/server

Yeah, Hetzner has 20 vCPU 64GB RAM i5-13500 servers for $40/mo:

https://www.hetzner.com/dedicated-rootserver/matrix-ex/

This is ~10x cheaper than the closest AWS option, and without the extra fees.

Also of note that Hetzner is profitable, which means AWS has been operating at an insane markup.

Exactly, it's the same with autonomous driving or any other activity (even art!) that has always required a human. While it can "look cool" and like it "understands the context" on the surface, when diving deeper it will _always_ misunderstand the nuance, because it doesn't have all of the lived human experience and reasoning.

Being even 80% (or 90 or even 95) there isn't enough - something will always be missed because it's only able to "reason" probabilistically within a narrow area not far away from the training data.

In principle, it's always been about "shared responsibility principle".

But in practice, yes many clients probably assume that by running in the cloud, they're "secure" (thus, failing their end of the bargain).

That's true - he was talking about clients though, if my memory serves well.

The main challenge he highlighted is there're no financial incentives for most companies in the industry to stay secure (unless you're a security company) - the punishment (including reputational risk) is just way too small.

FWIW from my own experience with auditors, the process is really kind of superficial. Yes, they can identify the most common checklist-based gaps, and that's what tends to be the low-hanging fruit for attackers as well. But they would never go deep enough to identify something that a determined attacker could exploit.