AWS has released aws-bench, an open-source benchmark for evaluating AI agents on real AWS tasks such as misconfigurations and infrastructure provisioning. Unlike traditional benchmarks, it uses real resources in disposable AWS accounts, scoring agent performance through automated verifiers. By Gianmarco Nalin
AWS has released aws-bench, an open-source benchmark for evaluating AI agents on real AWS tasks such as misconfigurations and infrastructure provisioning. Unlike traditional benchmarks, it uses real resources in disposable AWS accounts, scoring agent performance through automated verifiers. By Gianmarco Nalin
In context
- Topic: Desarrollo de Software — Arquitectura de aplicaciones, prácticas de entrega y calidad.
- Source: InfoQ
- Published: 22/08/2026
Continue reading at the original source →
Excerpt published automatically by the site radar. The full text belongs to its publisher and is linked above.
Why it matters
After a few years, nearly every organisation that migrated discovers the same thing: the bill grows faster than the business. Not because the cloud is expensive, but because nobody switches off what is no longer used and nobody has the incentive to check.
I separate what moves for cost from what moves for capability. Migrating to save money almost never works; migrating to be able to do something you could not do before does. When the argument is only savings, it is worth checking the numbers again because they are usually incomplete.
What usually goes wrong
Where it usually breaks is team capability. The platform is new but people learned on the fly, with no time and no support. They end up replicating data-centre practices in the cloud, which means paying cloud prices for benefits that never arrive.
What to watch
- How hard it would be to leave or move a piece to another provider, which is future negotiating power.
- Who sees the bill and in what detail: with no owner for the spend, the spend grows on its own.
- Where the data physically sits and what local regulation demands about that.
How I read this entry
If this fed into an infrastructure decision, I would ask for the three-year total cost with real growth built in, not year one with the entry discount. Most cloud surprises live in year two, when the discount ends and the volume has already gone up.
This entry is an excerpt from the original source, selected by the site radar. The commentary above is the site's own and does not belong to the cited publisher.
Living through this in your own team?
Open the chat and tell me how you're handling it. I'm interested in comparing notes.
