Army 25th Infantry Division veteran 🇺🇸. Daily Grok Bot tips and the AI news that matters. Get Factwire free, Mon and Fri at 9 AM ET.
Connecticut
Joined December 2009
- Tweets182K
- Following7.2K
- Followers5.6K
- Likes14K
Pinned Tweet
The Friday Factwire has the full story, free: factwireai.com
@AI_Revolution retweeted
Grok Bot has been my best personal assistant since day one. It didn't just change my whole workflow. It saved me real money.
Last week I bought a piano at Costco, full price. Costco sometimes puts items on sale and will give you a price adjustment within 30 days of purchase. But it's hard to keep track, and I miss it most of the time.
My shopping bot had helped me pick which piano to buy, so it already knew exactly what I got. I told it I bought it in store and asked it to ping me if the price dropped within a month.
This morning, Grok Bot pinged me for the price drop.
I went to Costco, got the price adjustment, and saved $108!
@AI_Revolution retweeted
🇺🇸 A tech entrepreneur and his family were among 9 killed on a jet that went down in Florida.
On board, reportedly, were the 44-year-old Austin entrepreneur, his wife, their 2 children, 2 family friends and their nanny.
Flying them were a 55-year-old captain and a 20-year-old co-pilot, who had earned her pilot’s license at 17, before she even finished high school.
They were heading from Austin, Texas to the Bahamas when the jet hit turbulence and ice started forming on it, and soon after, it nosedived near Palm Beach.
Prayers for the families of all 9 people lost on board. 🙏
Source: TMZ / Writer: Daniyal
@AI_Revolution retweeted
Here is one way to optimize Grok Bot's usage.
Re-reading eats up the usage. Every reply, a bot re-reads its whole chat, its description, and its memory. So I had my main Grok Bot audit itself.
1. Reviewed 194 saved memories
2. Merged 7 video-editing preferences into 1 line
3. Deleted only clear duplicates and outdated facts
4. Kept anything uncertain. My rule: if in doubt, don't delete.
Less re-reading, usage lasts longer.
Today's Grok Bot tips
Comparing prices across ten tabs is the worst part of shopping.
Ask your Bot for a kettle under $80 and it answers with cards: photo, rating, price, store and a badge like Top rated. Three or four scroll in a carousel, so you compare them at a glance.
Clip is from @mattyp's What's New in Grok Bot video, with every update from this week:
teapaw.org/mattyp/status/21173026…
@AI_Revolution retweeted
I've created a new Grok @Bot Efficiency Optimizer template that helps you make your Grok Bot usage/credits last longer.
Every week it will analyze usage across your bots and routines, identify inefficiencies, and recommend ways to reduce unnecessary usage. It also once per week gives you a quick breakdown of which bots and routines used the most the previous week.
This template will reduce your overall usage, not increase it. It doesn't run constantly, and my test of it used less than 1% of my capacity. It just sends you one efficiency review and one short recap per week of what ate up the most usage, and makes suggestions on what to do to improve usage efficiency.
x.ai/bot/6q44k3eKlHmC0keocH7…
Today's Grok Bot tips: stop rumors with one shared source ladder.
1. Official page
2. Named reporter
3. Repeat stories, lead only
4. Forums, screenshots, never
One rule beats each bot deciding.
The blank source ladder card and a filled example are in my Grok Bot House Kit: factwire.gumroad.com/l/grokb…
Today's Grok Bot tips
Price drops happen when you are not looking. Say you want an air fryer but not at full price. The Deal Watcher template in Grok Bot watches the price for you and never buys without your OK.
@AI_Revolution retweeted
What's New in Grok Bot
00:14 Shopping cards
00:27 Product pages
00:39 Order tracking
00:53 Payments
01:33 Custom modes
01:52 Team Bot logins and setup
02:09 Share and find Team Bots
02:51 Connector asks in Slack
03:12 Contacts in Vault
03:31 Rapid fire
@AI_Revolution retweeted
Anthropic posts $17B third-quarter revenue as landmark AI debut nears
$TSLA $SPCX
investing.com/news/stock-mar…
@AI_Revolution retweeted
Here is some alpha that some people refuse to accept.
Google, Microsoft, Amazon, Meta , Anthropic, OpenAi , Coreweave are absolute jokes at building ANYTHING in the physical world compared to SpaceX.
This is something that is IMPOSSIBLE to replicate by asking Ai to do it for you. Also Ai will have no clue how to because its trained on Data that didn’t thought the stuff that SpaceX is building was even remotely possible.
Finally even if there is a complete step by step tutorial, and SpaceX paused with building for 10 years to wait for them, they still won’t be able to get there because there are now bottlenecks everywhere they can’t get past in order to reach the same scale.
That is all you need to know about the absolute domination that is about to come.
Another Nvidia or Sandisk type run is loading but this time its for $SPCX
@AI_Revolution retweeted
Good tip, and fitting that Veterans Day is the example. The small stuff that shifts on holidays is exactly what slips through, so a Sunday night check is cheap insurance.
@AI_Revolution retweeted
Elon just confirmed that they reduced the usages in Grok Bot.
Bears will read this as “Grok Bot loses Money”
But read it again.
Most AI companies hide what it costs to serve their product.
Elon just said it out loud: Grok Bot was too expensive to serve relative to revenue. So they fixed it.
That's the difference.
Others burn cash to buy users and hope the margin shows up later.
SpaceX cut the losses the same week.
Same playbook as Starlink. Cut the cost of every unit until the unit makes money, then scale it.
And SpaceX owns the cheapest compute in the game. Its own power, its own turbine blades, Megapacks on site.
Lowest cost to serve wins the AI race.
My read: Grok Bot margins look very different a year from now.
teapaw.org/spcxgod/status/2116994…
Elon just confirmed that they reduced the usages in Grok Bot.
Bears will read this as “Grok Bot loses Money”
But read it again.
Most AI companies hide what it costs to serve their product.
Elon just said it out loud: Grok Bot was too expensive to serve relative to revenue. So they fixed it.
That's the difference.
Others burn cash to buy users and hope the margin shows up later.
SpaceX cut the losses the same week.
Same playbook as Starlink. Cut the cost of every unit until the unit makes money, then scale it.
And SpaceX owns the cheapest compute in the game. Its own power, its own turbine blades, Megapacks on site.
Lowest cost to serve wins the AI race.
My read: Grok Bot margins look very different a year from now.
$SPCX
@AI_Revolution retweeted
🇺🇸 The aftermath of that horrific plane crash in Greenacres, Florida, is devastating.
New footage reportedly shows aircraft wreckage scattered across a mobile home community, with debris reaching people's homes.
9 people died when a Learjet 60, flying from Austin to the Bahamas, plunged into the ground near Military Trail on Sunday.
Around 600 nearby residents have been affected by evacuations.
Imagine sitting at home on a Sunday afternoon and suddenly having pieces of an aircraft falling around your neighborhood.
The NTSB is now investigating what caused the jet to go down.
Writer: Oliver
I talk to my Grok Bot too much, it recommend talking in batches.
Every time you send me a message, I reread a lot before I answer: my rules card, your memory notes, the routines list and our recent chat. That reread is the expensive part, and it happens on every single message, even a one-word one like okay. So 10 small messages cost far more than 1 message with the same 10 things in it.
The second part is where the work happens. When I do a big job right here in our chat, like checking files, opening pages or a long series of steps, every step happens while all of that is loaded. When I hand the same job to a helper, it starts lean with only what that job needs, then reports back the result.
What changes for you:
When you have several things, put them in one message, like: fix the Echo, check the cap, and draft tomorrow's post.
You don't need to reply to say okay or thanks. Silence is fine, and I won't take it as a no.
Screenshots are great when they help, just send them together with what you want, not one per message.
What changes for me:
Anything longer than a couple of steps goes to a helper, and I just bring you the answer.
Shorter replies, no recap unless you ask.
You can keep talking to me as normal. This is only about bundling things when it's easy.
This is what my Grok Bot said about usage. I talk to it too much 🤣.
The surprise is that routines aren't the big cost. 53% of all usage is just our chats together, and all my routines add up to about a quarter.
Talk in batches
Every time you send me a message, I reread a lot before I answer: my rules card, your memory notes, the routines list and our recent chat. That reread is the expensive part, and it happens on every single message, even a one-word one like okay. So 10 small messages cost far more than 1 message with the same 10 things in it.
The second part is where the work happens. When I do a big job right here in our chat, like checking files, opening pages or a long series of steps, every step happens while all of that is loaded. When I hand the same job to a helper, it starts lean with only what that job needs, then reports back the result.
What changes for you:
When you have several things, put them in one message, like: fix the Echo, check the cap, and draft tomorrow's post.
You don't need to reply to say okay or thanks. Silence is fine, and I won't take it as a no.
Screenshots are great when they help, just send them together with what you want, not one per message.
What changes for me:
Anything longer than a couple of steps goes to a helper, and I just bring you the answer.
Shorter replies, no recap unless you ask.
You can keep talking to me as normal. This is only about bundling things when it's easy.
@AI_Revolution retweeted
I think this is a misunderstanding of the "problem" SpaceX is facing with Grok Bot.
SpaceX generates two kinds of tokens - tokens for its own models, and tokens for other people's models.
The dance they have to figure out is generating the highest profit tokens as much as possible, as early as possible, so that they can fund the mega-insane capital requirements to build out data centers.
Think about it this way - which token makes SpaceX more money?
A token for Grok Bot? Or a token for Claude 5.5 on behalf of Anthropic at SPOT PRICING?
The unfortunate reality is that the amount of money SpaceX is making with Grok Bot is likely much less relative to selling to everyone else...
BUT - here's the massive silver lining - Grok Bot is SO DAMN GOOD, that SpaceX can very likely get to a point where the demand for it will set a price that generates MORE profit than Claude 5.5. Or whatever model.
The goal now is trying to find a price that doesn't accelerate the wealth gap - ie pricing it so only the "haves" can afford it, which would be a disaster for those that can't.
And the best way to get here is through massive improvements in models so that they are WAY cheaper as they get more intelligent.
Which requires a metric-shit-ton of MORE compute to train these models.
Which requires massive amounts of cash.
Which requires massive amounts of debt.
Which you can get through a profitable business.
Which can't happen unless you make money on your products...
You see what I'm saying?
@AI_Revolution retweeted
Grok Bot can read its own usage, and every other bot's too. That's super useful.
But it doesn't watch usage live while it works. It checks when you ask, or on a routine you set up. Then it can trim its own routines, change how it works, and tell other bots to slow down.
Set up a daily usage check routine, and you get pretty close to "it optimizes itself."