Terry Evans (@lordsnave) header
Terry Evans (@lordsnave) profile picture

Terry Evans

@lordsnave
Joined April 2009
  • Tweets2.1K
  • Following119
  • Followers995
181Media
Photo by Terry Evans (@lordsnave)Photo by Terry Evans (@lordsnave)Photo by Terry Evans (@lordsnave)Photo by Terry Evans (@lordsnave)Photo by Terry Evans (@lordsnave)Photo by Terry Evans (@lordsnave)Photo by Terry Evans (@lordsnave)Photo by Terry Evans (@lordsnave)Photo by Terry Evans (@lordsnave)Photo by Terry Evans (@lordsnave)
Why is the voice chat integration with @bot so horrible? @poteto I enter voice mode, talk for a minute or two, with the AI responding to me the entire time, but none, or very little of the actual conversation make it to the bot itself - it has no record of what was said.
49
When someone is essentially threatening to throw you in jail when they get the power to do so, you’re going to do everything possible to make sure they are never in power again. These people are very dumb.
Elon Musk (@elonmusk) profile pictureElon Musk
@elonmuskOct 21
Replying to @alx
Message received
55
All this hatred towards Tim Sweeney and using AI for games is giving me the opportunity to mute a lot of people. Use AI to enhance your craft, but this whining about it is just annoying.
3
91
$TSLA Roadster unveiling on the 15th! youtu.be/JFT7hNhop7w
When This Baby Hits 88 MPH ... Back to the Future Quote

When This Baby Hits 88 MPH ... Back to the Future Quote

When This Baby Hits 88 MPH ... Back to the Future Quote

youtube.com
Christopher Lloyd (@DocBrownLloyd) profile pictureChristopher Lloyd
@DocBrownLloydOct 13
Almost time to see some serious shit…
74
Anyone else getting annoyed by all the @bot tags in post replies that just add spam and no value to anyone except that one person who tagged their Grok Bot? @poteto what you think?
70
Bye. @OpenAI @thsottiaux
Image from Terry Evans (@lordsnave)
Terry Evans (@lordsnave) profile pictureTerry Evans@lordsnaveSep 29
Replying to @thsottiaux
"Today we are going to ship a number of things that increase what you can do across the Plus and Pro plans." I use your plans to code. I don't care about extra fluffy features I'll likely never use. If the ability to code is nerfed, I will cancel, plain and simple.
81
I've mentioned frequency about how bad Claude's subagent system is, so I've found my solution. I'm using OpenCodex for routing. Then using Grok Build as my main CLI and selecting whichever model I want from the OpenCodex model list. So far, it's working a lot better.
1
2
110
You can get OpenCodex here: github.com/lidge-jun/opencod…
GitHub - lidge-jun/opencodex: Universal provider proxy for OpenAI Codex & Claude Code — use any LLM...

GitHub - lidge-jun/opencodex: Universal provider proxy for OpenAI Codex & Claude Code — use any LLM...

Universal provider proxy for OpenAI Codex & Claude Code — use any LLM (Claude, Gemini, Grok, DeepSeek, Ollama…) with Codex CLI, App, SDK, and Claude Code - lidge-jun/opencodex

github.com
47
Does anyone even use the plan mode any longer for AI agents? I still create plans (from a spec), but never use the actual plan modes. I haven't in a couple of months now.
100
I'm never going to use Astra to execute an implementation plan again. It's so horribly slow at it. Like, really, really bad. Then it just feeds you one bs excuse after another if you ask why it's so slow.
122
Claude's auto-classifier is so stupid I had to use Grok to fix Claude. Boy, they make some dumb decisions at Anthropic.
2
122
One thing @herdrdev needs is the ability to search for text within a tab. 🙏
108
Astra's image generation kind of sucks when you are trying to iterate on an image/design. It keeps drifting each image instead of starting with the original and making just the deltas you want. They need to fix that. @OpenAI
1
110
The main issues I saw: Image looses details over time as it iterates on an image and not the actual design description Changing a specific color changes ALL the colors so the one color you want will match Forgets what the original design even was so just removes stuff you want.
1
31
Something like this would be far better implemented by keeping the original design description and altering it to create a new, updated image each time instead of trying to iterate on a screenshot of the design each time. Keep the original design image as a reference for it.
13
I just cancelled my $100/month SuperGrok Plus plan. It just isn't worth it at this time. I'll try again in a few months. That money is better spend to use Fable or Astra.
Elon Musk (@elonmusk) profile pictureElon Musk
@elonmuskSep 14
Replying to @itslueul
Grok 4.7 should be roughly on par with Opus 5.0, not 5.1. Better in some ways, worse in others. We need to fix multimodal performance. Grok 4.8 will be a noticeable improvement. Grok 4.9 is probably Astra/Fable class. Grok 5 maybe better than anything. We shall see.
1
128
Well, @claudeai @AnthropicAI 's subagents are still horribly implemented. I get so frusrated each time I try and use them. So incredibly unreliable. What a waste of time and tokens.
214
I started using herdr earlier today, and I like it so far, but I did encounter a nasty bug: github.com/herdrdev/herdr/is…
Windows 0.9.0: Local server disappears and every TUI window stays on reconnecting · Issue #3858 ·...

Windows 0.9.0: Local server disappears and every TUI window stays on reconnecting · Issue #3858 ·...

Is this a reproducible bug? I confirm this is a reproducible bug, not a feature request, idea, question, contribution proposal, or direction check. I reproduced this bug on the version and environm...

github.com
140
I had to have @grok fix the @OpenAI Codex install issue I was having. 🤷‍♂️
1
151
I wish I could set a default effort level per AI model. That way I could have a default like Fable/low and Opus/high, and when I switched between models I would get those effort labels autmatically.
132
Load more