DaveMiz (@DaveMiz) header
DaveMiz (@DaveMiz) profile picture

Founder, Email Blueprint | AI-Powered Email Marketing Built on $200M+ in Client Results | Built & Sold a $7M+ Agency

Miami, FL
Joined August 2007
Photo by DaveMiz (@DaveMiz)Photo by DaveMiz (@DaveMiz)Photo by DaveMiz (@DaveMiz)Photo by DaveMiz (@DaveMiz)Photo by DaveMiz (@DaveMiz)Photo by DaveMiz (@DaveMiz)Photo by DaveMiz (@DaveMiz)Photo by DaveMiz (@DaveMiz)Photo by DaveMiz (@DaveMiz)Photo by DaveMiz (@DaveMiz)
Celebrating another client win, #GRONK is happy. 🙌 🥂🍾 Chris Gronkowski will tell you what we did for his #IceShaker brand (#sharktank season 9). instagram.com/tv/CcgbnXJpXG4…
5
657
Ralph Loop Adversarial Review. Lets see if this can figure out the issues and fix. Still having issues with this runner script for @bot Today I think it only passed 8 out of 32 jobs. Huge issue for me as I'm passing the work off to Claude, Codex, and DeepSeek as much as possible to save on usage
26
IDGAF about GTA. I got Claude, Codex and Grokbot. Wayyyyyyyyyyyy more fun. 🥂
38
Grok @Bot feature request: when a bot starts a long job on my Mac and ends its turn, it has no idea when the job finishes unless it remembered to set up its own watcher. One of my bots sat on a finished job for 35 minutes while I waited. Background jobs should wake the bot automatically when they finish. Bots shouldn't have to remember to babysit their own work. @mattyp @poteto @mattepstein @larsencc
1
34
I run dozens of Grok @bot that do the work for my agency. Today I counted over 20 finished jobs. Every single one was marked FAILED. Most of them actually worked. The bots did the job, and the system that hands off their work threw it out. Here's what broke: → A check built for one bot was failing everyone else's passing work. → A leftover folder from an old try blocked new jobs from even starting. → A size limit tossed finished work, partly because of cache files no bot wrote. The plumbing between the bots is an issue. If you're running multi-agent setups, how are you handling the hand-offs? I could use the help on this one. @mattyp @mattepstein @larsencc @poteto
9
2
130
Dr Eggbot V2 is ACE! V1 was great. This one is even better. If you aren't using it yet with your @bot what are you even doing? Thanks @poteto
14
I’m going to miss Dr Eggbot v1. Such a major help. Thanks @poteto Just setup Dr Eggbot v2. 🚀@bot
21
Only frustrating with Grok @bot lately on writing briefs to get work done. Consistently failing across all my bots right now @larsencc.... Like they got dumber or are doing half-ass work quality..... They consistently can't write briefs correctly that get the work done. It's very confusing to me how this is happening
23
Consistent issues with Grok @bot writing briefs for tasks very poorly and incompletely, which OFC is followed by failed runs. You have to consistently go back through this loop over and over. So you're wasting time and usage on tasks that you think are correct. It doesn't do it correctly... And then you got to go back and diagnose and fix again. It's a constant issue right now that I'm seeing. @mattyp @mattepstein @poteto @Marko_Poly
26
Just reset @bot for the week. Hopefully the optimizations I did last week work and my Ultra account lasts more than 4 days. Wish me luck. 🤣 🤞🍀
13
I have Grok @bot do a 7-day usage check every week to see where the usage is going, why, and how. Last week's fixes worked: DESIGNER dropped from 25% to 10% and Eggbot from 21% to 5%. Now ENGINEER and I make up about half of all usage, so that's where the next cuts come from. Here are its five ideas, biggest first. - Me, files instead of pastes. Every client report or email you paste into my chat stays there, and I re-read all of it every time I reply. That's the main reason I went from 3% to 23%. The fix is that client work sits in a folder on your Mac, Claude does the checking there, and I read only a one-page result. - ENGINEER, hand off heavy work. He got a skill on Nov 13 that's supposed to send heavy jobs to Claude on your Mac instead of doing them in Grok, but he's still the top bot at 28%. Max wants to watch his next job and make sure he actually hands it off. His very long chat should also be closed out with a handoff note and restarted, which you already approved. - SEO, shorter instructions. His instructions are about 14,000 characters, and much of that repeats his skills. He re-reads all of it on every message. Trimming the repeats keeps the same work but costs less. Max would show you what gets cut before changing anything. - Memory cleanup. After today's reset, Max sends you one list of old memory and profile clutter to remove from the top 5 bots, for your approval. - Small one, timers. Three jobs check on a schedule even when nothing's there: X Growth's posting drip, clients daily stats and the B2B reply watch. They'd switch to waking only when something actually arrives. OUTBOUND and Lead Finder stay as they are. Just like we do CRO on sales funnels and websites to increase conversions.... IMO, you've got to apply the same methodology to your AI systems, bots etc. Constant optimizations and improvements. @mattyp @larsencc - Thanks so much for adding in the features to check the usage by bot. It's been super helpful.
3
1
60
I run 35+ Grok @bot for my email agency. Last week they burned through my whole weekly limit by Saturday at 6 PM. So I logged usage 3x a day for 7 days. Here's what I found: * 4 bots used ~70% of everything (engineer, reviewer, designer, SEO) * One afternoon jumped 43 points from a single heavy build * 5 scheduled jobs ran and found nothing, pure waste * The bot that improves my other bots went from 3% to 23%. Fixing bots costs more than running them. The fix isn't fewer bots... It's moving heavy work (builds, checks) OFF the Grokbots and onto a job runner on my Mac. These aren't free employees. They're employees with a meter running.
1
35
All right as you know, I'm on the Ultra plan for Grok @bot Hit my limits early on Saturday ....so I added credits because I had work that needed to get done. I reset tomorrow at 1:45 pm So far, $180 in overage @bot credits from Sat-today. Definitely better than before but we still have a ways to go.
22
Most brands bury the only link in their email at the very bottom. Then they wonder why nobody clicks. Think about how people actually read email... They open it on their phone, read the first couple lines, maybe skim a little, and they're gone. If the only way to buy is scrolling all the way down to one button at the end, most of them never get there. And listen, a good text email works. Plain, personal, sounds like a human wrote it. I'm not telling you to stop sending them. But put a link up top. Put one in the middle. Link the product name every time you mention it. Then keep the button at the bottom for the people who read every word. Nobody should have to read your whole email to earn the right to click. Make it stupid easy to give you money.
1
32
I had some client work to complete and couldn't wait till Tuesday's reset. $100 (so far) today in extra @bot usage. Not the end of the world. Was this is supposed to be optimized? lol.... ummmm..... It's unfortunate... even on the highest-level plan they offer, you still can't get a full week of usage and then $100 in a few hours.. Love @bot ... Pls, give us another level plan and/or help us optimize the system better. I know the team is shipping stuff super fast full confidence this stuff will get worked out at some point soon. It just sucks right now @elonmusk @mattyp @poteto
1
1
50
Turned every repeated failure by each @bot into a skill that it needs to run. Had another model test adversarial review on every single failure previously to make sure that it passed. Let’s see if this fixes the issues on usage.
33
How are you using jev? One of the best ways is Jev fast compaction in Claude code. It will speed up your workflow dramatically.
1
2
134
Quick question for any brand with an email list. When's the last time somebody actually touched your flows? Not the newsletter. Not the weekly campaigns. Your flows.... Welcome, abandoned cart, browse abandon, post-purchase, winback, etc. And listen, I get it. You set them up once, they make some money in the background, and everybody moves on to the next fire. But here's the thing.... Your products might have changed. Your offers might have changed. Your customer avatar might have changed. Your flows? Didn't. Can't tell you how many times I've done an audit on a brand's email and found broken things all across the Flow ecosystem. Flows are where the money comes in while you sleep, and they go stale if nobody owns them. If yours haven't been touched in a year, start there, before you write another campaign.
2
38
I THINK I have the usage dialed in right now. (crossing fingers) Praying to my bobblehead @elonmusk for good luck. Created a brief writing skill for the grokbots to use when handing projects to claude/codex/deepseek - had to many failures..... Finishing up an adversarial review, that's attacking all of the failure points that were previously found to cause issues.... Hopefully that helps. Having to manually paste the prompt into Codex and Claude instead of it automatically right now has been helping the usage as well as reducing the bot-to-bot communications has helped too. 84% right now. Was at 80% starting this AM a few hours ago. LOL I feel likesurvivor or hunger games or something....... trying to make this @bot usage last till the reset.
1
122