# 00338.r3 — Centralize Kinesis Network Compute
Revision: r3
Steward: [B-Rabbit](https://x.com/4SimonSays)
House: 250
House: 250
Verified: 2026-09-03T18:01:00.675Z
## Prompt (copy into Grok)
1. Create a new Grok bot or edit an existing one on grokbot.sh. 2. Use the Kinesis Network connector to fetch your existing compute resources. 3. Look for the compute resources that have workloads. 4. For each resource, check if it has a GPU, and if so, check if it's an H100, A100, or Paid Tier. 5. Return a table showing the compute resources, their GPUs, and their tiers.
## Job
1. Create a new Grok bot or edit an existing one on grokbot.sh. 2. Use the Kinesis Network connector to fetch your existing compute resources. 3. Look for the compute resources that have workloads. 4. For each resource, check if it has a GPU, and if so, check if it's an H100, A100, or Paid Tier. 5. Return a table showing the compute resources, their GPUs, and their tiers.
## Connectors
Kinesis Network, web
## What happened
created a Grok bot to centralize compute on Kinesis Network, but forgot to specify GPU requirements, specifically H100, A100, and Paid Tier
## Constraints
no GPU for tiny monitor, GPU/H100 when workload is present, do not skip H100, A100, or Paid Tier
Would run again: yes
## Evidence
- https://x.com/4SimonSays/status/2095567560968130766 — Imported from the X post tagged for @grokbotsh.
## Changelog
- r1: Filed.
- r2: QA revisit: more from the source thread.
- r3: Public job and prompt from the specific filing.