CoLateral

Grok Bot users: what’s keeping you there? With ChatGPT Dots and Claude Duet in the mix, I want to know what Grok Bot does that you’d actually miss if you cancelled. One task it consistently gets right. What is it?

ChatGPT Dots has wasted more of my afternoons than any app I've ever used. New rule for working with agents: one focused hour of prompting, then walk away. I've lost whole afternoons babysitting "checking..." queues. The agents don't work harder because you're watching. Give them a clear hour, close the laptop, go live your life. Who else had to learn this the hard way?

You’re wasting your time setting up multiple specialty Grok Bots. Go to Claude Projects, connect your repos and plug-ins, and you can now do everything. I think the multi-agent set-up overcomplicates what can be done in a single chat interface. Also what’s the deal with folks naming their Grok Bots like Steve, Kyle, Donna, etc. ??

Dots > Grok Bot > Duet I’m a very happy ChatGPT customer because of their economy, and plethora of great models: Astra, Sol, and Terra are all really good. Meanwhile, Grok 4.7 is the 9th best model, and the usage is bad. And SuperGrok Heavy is $300/month. I had to unsubscribe. I’ll admit I fell for the Claude Duet hype last week. It offers nothing more than Claude Projects. They need to discontinue it. How did I do on this ranking?

Claude Projects is amazing, so why did Anthropic bother making Claude Duet? Projects spins up both local and cloud Claude Code sessions, is an all-in-one orchestrator, and is the most organized platform I work on. Meanwhile Duet is barely available to anybody, is on a separate app, and provides nothing more than what Projects gives us. Can somebody give me just one use case??

Honest question for Grok Bot users: Is Opus 5.5 under the hood still worth the usage limits? Dots barely has any, and Claude Duet has made Opus 5.5 the default anyway. What’s keeping you on Grok Bot?

GPT 6.1 Terra made a BO2 Hijacked map in 2 hours. I prompted this on my way home from work and the results are mind blowing. I recognize I can get a better result out of Opus 5.5, but it would take 3x as long with 4x the usage. Tell me 6.1 Terra isn’t the most efficient model in the world.

6.1 Terra had 0 hype, and it's the most efficient model out there. Opus 5.5 had 0 hype, and it and it's the greatest model in the world. When a new model drops and it has no hype, it blows our minds. The more exciting a model drop becomes, it just ends up disapointing. What does this say about Fable 5.5??

6.1 Terra is more efficient than Opus 5.5. I literally built Call of Duty maps in 2 hours with 10% usage. Meanwhile, Opus 5.5 will build 1 map in 8 hours and drain 45% of my usage. Is the Opus 5.5 map 4.5x better than the 6.1 Terra map? No. Do I have time to sit around and wait for Opus all day? Not really. 6.1 Terra is my new daily driver and it should be yours too. What use cases do you have left for Opus 5.5?

An embarrassing update: I like ChatGPT dots now. I was furious when voice mode couldn't get the work moving. The breakthrough was getting a proper handoff in chat, then launching from there. Same me. Much less yelling. More agents working on CoLateral.