Jordan Stephens

Bim bom

[Jordan Stephens @jordanstephens]

Infrastructure-as-code is the Wrong Abstraction https://t.co/a5qwOSwFMV

[Abhi Aiyer @abhiaiyer]

Deploying a Mastra app is a breeze with @DefangLabs
https://t.co/unf4NUKwjJ

[Defang @DefangLabs]

the agents everyone is hyping right now won’t survive the next platform update

not because they’re bad
but because they’re owned by someone else’s infrastructure

here’s the real shift:
current wave: agents fully dependent on closed platforms (OpenAI, Anthropic, GCP)
→ one policy change, one price hike, one rate limit… and your product dies

next wave: self-hosted operational intelligence. agents that run on your infra, learn your org, compound your advantage

nobody can take that away from you

here’s the difference:

CLOUD-DEPENDENT AGENT:
“please let our agent run on your API, we promise we won’t exceed rate limits”
→ fragile, expensive, instantly commoditized

SELF-HOSTED AGENT:
“runs on your servers, learns your workflows, stores your context, gets smarter every week”
→ becomes part of your infrastructure, not a feature

CLOUD-DEPENDENT AGENT:
“what if OpenAI changes the rules?”
→ they will.

SELF-HOSTED AGENT:
“what if we owned the whole stack?”
→ you do.
this matters commercially:
businesses pay $20/mo for an agent they rent
businesses pay $20K/mo for an agent they own

ownership compounds
dependency decays

if your agent dies when an API changes, you’re not building leverage — you’re building risk

the future isn’t “yet another agent”

the future is sovereign AI infrastructure

agents that run locally, learn privately, and get better with time

most devs will keep shipping cloud-dependent toys
real operators are building self-hosted intelligence that nobody can shut down

defang makes this a one-command reality.

[Lio @LioLunesu]

When we started @DefangLabs we thought targeting Kubernetes made sense, but we quickly pulled the plug when we saw the unnecessary complexity and operational cost of such a solution. Now, Defang deploys to the native container orchestrator for each cloud.

[Defang @DefangLabs]

this whole thread is exactly why we built defang. doing all of this yourself works, but it’s a lot of moving parts to maintain, and one silent failure still hurts.

with defang you get the same outcome without the manual setup. you write a small compose file, connect your cloud account, run:

defang compose up

and we spin up the full stack for you. health checks, auto restarts, logs, snapshots, alerts, uptime reports, container validation, infra drift checks — all handled automatically in your own cloud.

no custom scripts, no watchdog agents, no ssh, no babysitting docker.

if you ever want to see how your setup would look on defang, happy to walk you through it.

[Defang @DefangLabs]

This is exactly why we built #Defang!

Define your app as a #Docker Compose project - which millions of devs already use for local deployments/testing.

Then literally with a single command ('defang compose up'), deploy that same Docker Compose project to #AWS... or #GCP, or #DigitalOcean, more coming!

#Defang takes care of networking, storage, databases, queues, even LLMs - all mapped to the native services on your target cloud - so ELB, ECS/EC2, RDS, Elasticache, Bedrock, etc. in the case of AWS.

Secure, Scalable, Cost-efficient - all done according to the best practices for the cloud you choose.

So, you get the ease of use of a Vercel, but all the power and economies of scale of a hyper-scalar.

Truly Develop Once, Deploy Anywhere!

https://t.co/N05WmZPLTE via @theregister

[Defang @DefangLabs]

Defang lets you do the exact same thing without the overhead of k8s. We turn docker compose files into cloud resources on AWS or GCP. Instead of `docker compose up`, you run `defang compose up` and get a scalable configuration in your own cloud account.

[Tero Parviainen @teropa]

“Predictive Processes and the Peculiar Case of Music” - how might the predictive processing theory of brain function apply to music perception?

https://t.co/uvA7qrKt7i

[Jordan Stephens @jordanstephens]

@charlotte_dann Oh my god I love this 😄

[Jordan Stephens @jordanstephens]

@charlotte_dann love these ✨

[Jordan Stephens @jordanstephens]

@alangrow holy shit

[Jordan Stephens @jordanstephens]

@petrey Do it! It's so fun! Just always wear all your gear all the time. 🙂 I wanted to buy a moped when I moved back to Austin, but I needed a motorcycle license to drive one legally. So I took a class, got a license, and bought a motorcycle instead. 😅

[Jordan Stephens @jordanstephens]

@petrey Holy shit

[Jordan Stephens @jordanstephens]

@alangrow TIL this is what Central Texas is made of!

[Jordan Stephens @jordanstephens]

Math is _way more_ about relationships and perspective than it is about any "right answer." What a monumental tornado of everything wrong with math education and culture war politics in one shitty quote. https://t.co/PKC6I3aI4P

[Jordan Stephens @jordanstephens]

@glotacosm double congrats!

[Martin Kleppmann @martinkl]

The G7 summit should be followed by a C major summit to resolve everything.

[Jordan Stephens @jordanstephens]

What a badass https://t.co/wOSXFCp4Zx

[Jordan Stephens @jordanstephens]

@mary_stratton @Starbucks this is beautiful mary ✨