ry

8.3K posts

ry banner
ry

ry

@rywalker

CEO @Tembo → Move engineering work to the cloud

Katılım Mayıs 2008
1.9K Takip Edilen12.3K Takipçiler
Sabitlenmiş Tweet
ry
ry@rywalker·
Proud of the team. The new Tembo reflects how far we've come. More good things ahead as we launch exciting new features. We're building the infrastructure so engineering teams can more easily put real work in front of any coding agent — and get merge-able production code back. Cloud or self-hosted. Your agents. Your control.
Tembo@tembo

New look. Same mission. Move engineering work to the cloud.

English
2
3
16
957
Mike Miller
Mike Miller@mikemillercyber·
Serious question from a founder. Do we even need business cards anymore?
English
9
0
9
2.5K
bruno
bruno@tvnxty·
I remade sound design for @tembo from scratch! Sound on! 🔊 Built from the ground up, more energy, more cohesion and audio that actually flows with the visuals, not as a secondary, but as an equal. Visual overhauls should be exciting, and what better way to make them so, than with detailed audio that complements your visuals. Motion is 50% audio, and no serious brands should neglect it. What do you think? Need peak audio for your brand? Let's talk, DMs open!
Tembo@tembo

New look. Same mission. Move engineering work to the cloud.

English
7
2
29
3.3K
ry retweetledi
ry
ry@rywalker·
Proud of the team. The new Tembo reflects how far we've come. More good things ahead as we launch exciting new features. We're building the infrastructure so engineering teams can more easily put real work in front of any coding agent — and get merge-able production code back. Cloud or self-hosted. Your agents. Your control.
Tembo@tembo

New look. Same mission. Move engineering work to the cloud.

English
2
3
16
957
James Falconi
James Falconi@Joi2James·
I wanna be your first customer Share your startup in one sentence
English
148
3
84
9K
Tembo
Tembo@tembo·
New look. Same mission. Move engineering work to the cloud.
English
50
17
336
46.8K
ry
ry@rywalker·
@dominikkoch @tembo when you team up w/ cracked minecraft kids... and they finally express their personality into the brand...
GIF
English
1
0
2
57
Dominik Koch
Dominik Koch@dominikkoch·
@tembo Why does the logo look like the minecraft creeper?
English
2
0
12
287
ry
ry@rywalker·
@milldr_ @tembo I hear something about snapshots is in the forecast?
English
0
0
2
24
ry
ry@rywalker·
@johnyeo_ at least a customer didn't churn :)
English
1
0
1
156
John Yeo
John Yeo@johnyeo_·
had bird shit fall on my laptop while doing customer support, just another day as a founder
John Yeo tweet media
English
21
0
79
3.6K
ry
ry@rywalker·
I keep seeing the same architecture work across very different problems. It is not a mega-agent. It is not a chatbot taped to an API. It is a fabric of small, specialized programs — each one doing exactly one thing with high confidence — coordinating through shared context. The mega-agent fantasy is seductive. One model, one prompt, one black box that takes your question and produces an answer. It works in a demo. It collapses in production the first time it has to do anything that matters because you cannot tell which step it got wrong, and you cannot trust the output enough to act on it. The mesh is the opposite. An enrichment agent qualifies an inbound visitor. A routing agent decides where they go next. A content agent rebuilds the experience around them. An analytics agent watches all of them and feeds learnings back. Each one is small enough to inspect, narrow enough to be confident, and replaceable on its own. The same shape applies to operations — a data-collection agent, a prioritization agent, a delivery agent, a capture agent. Different domain, identical pattern. What makes the mesh work is not the agents. It is the discipline around them. Context flows through written state. Each agent observes, applies its rule, hands off. Humans review the outputs that change the strategy, not every keystroke. The agents do the integration work. The humans do the judgment work. I've argued elsewhere that the ERP is dead and that the gap between ad click and revenue is a mesh problem. The mesh is the architecture under both. Once you see it, you cannot unsee it — every enterprise problem worth automating has the same shape. rywalker.com/mesh-of-specia…
English
3
0
22
2K
ry
ry@rywalker·
@aar0ncpa i think it’s both - as we optimize token spend we’re going to want a data cache (db/warehouse) as part of that optimization
English
1
0
1
35
Aar0n
Aar0n@aar0ncpa·
@rywalker I think a solution challenge is as follows: 1. Should I move all data to a data warehouse? Or 2. Should I give agents MCP access to various systems? While at the same time building an enterprise second brain for the agents. Being able to frame a solution here is challenging.
English
2
0
1
49
ry
ry@rywalker·
The starting point for enterprise agents isn't automating one process. It's replacing the tab-switching tax across six systems with a single execution layer that serves up priorities, helps execute the work.
English
5
0
13
731
ry
ry@rywalker·
Dream together, not solo. Share the map, carry each other through storms, and celebrate milestones. The journey feels lighter, richer, and far more meaningful.
English
0
1
6
273
ry
ry@rywalker·
The biggest barrier to enterprise agent adoption isn't the technology. It's inertia. Your PMs will tag an engineer before they tag an agent every single time, until someone forces the habit change.
English
1
0
8
643