Teacher → Tech. I test AI tools so you don’t have to. I train teams to get the most out of their technology.
Casper, WY
- Tweets 9,848
- Following 825
- Followers 3,795
- Likes 52,900
Pinned Tweet
This is a real party, and I’ll be there @sama
I’m not getting a Goblin tattoo though
GPT-5.5 is going to have a party for itself. it chose 5/5 at 5:55 pm for the date and time.
if you'd like to come, let us know here: luma.com/5.5
codex will help the team pick people from the replies. 5.5 had some good ideas/requests for the party, which we'll do.
8
3
62
22,345
this struck a nerve with many...and here's why:
there are 2 things millennials universally hate doing:
1. talking on the phone
2. answering the door
why are we recreating one of these with such a future oriented product? @sama
my @ChatGPT Dot took 5 rings to answer me
my @grok @bot answered in about .5 seconds, no fake phone UI or ringing
this is a terrible UX @thsottiaux , no one wants to pretend we are on more phone calls in 2026
3
3
394
it’s clear that the biggest thing holding @SpaceXAI back is the quality of @Grok models. they are not reliable for long running, complex, design focused, or writing tasks
if/when this is addressed, it will be the best sub and remove the need to delegate so much work to the more powerful models from other labs
combine a powerful Grok/Composer model with what’s great in their ecosystem, and it will be amazing
-@bot is best assistant app for real work
-@cursor_ai mobile apps and cloud agents are amazing
-grok build harness is best TUI/UX
-rate limits on 1st party models are solid
7
1
13
1,164
think of @NousResearch Hermes as OpenClaw...but it works
1
4
472
this is the reason why @ChatGPT plans are such a poor value since the changes:
1. the only usable model by rate limits is 6.1 Sol
2. it's a solid model, feels like Astra lite
3. but it's sooooo slooooow, this simple task took 9 min
so yes, you get a ton of usage on your plan with this model, but it's unusable for anything timely or multitasking...which is most of the time for most people
179
me before using @herdrdev: "why would i want to use a terminal?"
me after using herdr...
working across local and VM's all on my tailnet through one interface is amazing
3
405
my @ChatGPT Dot took 5 rings to answer me
my @grok @bot answered in about .5 seconds, no fake phone UI or ringing
this is a terrible UX @thsottiaux , no one wants to pretend we are on more phone calls in 2026
114
24
988
83,547
Dots have to be the software only precursor to a physical “dot” like these
I’m thinking Google Home + VM + connected to your @ChatGPT account and context
4
291
we are all creative directors now
this is @claudeai desktop clipping from a long form recording with @ryan_doser13, adding graphics and captions, naming the clips, then getting feedback from me
all in one app, no more switching windows to finder and launching videos 1 at a time
2
2
5
447
one thing no one talks about with @claudeai models:
they are exceptional at coordinating other agents
ex: Fable 5.1 in the cloud is delegating to @cursor_ai agents through @linear and tasks are synced between them within 1-2 minutes, with no interventions from me
3
5
579
what's different when you designate a "primary @bot"?
1- a start icon is added to it
2- it gets some sort of system-level instructions to be more proactive
example: I already had a Chief of Staff @grok Bot, but designated it at my primary
*you get a nudge to do this when opening the desktop app
it immediately started looking across my project folders and tasks for things it could help without me having told it to
sounds very similar to the @chatgpt Dot set up instructions for a "heartbeat dreamer" style agent, which you can port into your Grok Bots here:
x.ai/bot/mDdDU-Ex2yDqySZ5hiT…
8
1
39
5,197
teacher: “how did you ace your spelling tests every week?”
9 yr old son: “my dad and I vibecoded a web app to practice them each week in multiple modalities, track my progress, and use spaced repetition to review based on my performance”
2
1
11
437
Welcome to our live show! x.lingyaoai.com/i/broadcasts/1yGBePAjq…
2
1
5
524