Aaron Makelky
@theaaron
Teacher → Tech. I test AI tools so you don’t have to.
I train teams to get the most out of their technology.
784 Following 3.7K Followers
how tf is GPT 5.5 the top model… ?
if you tried hyperframes, you would know
the agentic video stack is being built on code-gen
but SWE-bench code v.s. code-to-video that feels alive are two different things
we built the Code2Video Bench, in collab w/ Google DeepMind & @Kaggle
frontier labs can finally get good at agentic video tasks
Show more
keeping track of which tool is working on which part of which project is a pain
embarrassed that it took me this long to follow @altryne's example & use @linear as the agent coordination tool
Show more
a year ago, we thought copy + pasting text was amazing from AI
now, i wake up to emails from my coding agent that set up a landing page to include a meeting link, then tested it out by booking me with it, then cancelled them so they don't stay on the calendar
Show more
use the in-app browser in @ChatGPT for 1 week, you'll use Chrome 90% less
this is @OpenAIDevs Codex + @linear in app browser side-by-side
i don't want your agent in my browser via an extension, i want my browser in your agent
Show more
for those using @typesafeai Jev to filter emails, here's a great use-case:
when the signed named doesn't match the inbox name= SPAM
this makes no sense to compare to Fable 5.1 Max and GPT 5.6 Sol Max
where is the GPT 6 Astra comparison?
Grok 4.7 is here.
It's a notable improvement over Grok 4.6 at the same price and speed.
Try putting the same amount of $ from one of your AI subscriptions on @NousResearch portal and use GLM 5.3 flash for a month
You’ll get close to frontier intelligence for pennies per task
Open Source AI Models are Closing the Gap
My latest video with @theaaron breaks down:
- What tasks are best for using open source models
- The best way to use open source models (through @OpenRouter)
- Our favorite open source models
- The $100 subscription math that kills the case for buying your own GPUs
- Where open source still falls short against frontier models
Watch the full episode on YouTube below 🤝
Show more
this is killing me tonight @bot
the only way to make Siri useful:
Settings -> Apple Intelligence & Siri -> Extension -> @ChatGPT
when these talk to @Muse … wow, they will go from novelty to best application of consumer AI over night
these don't need to exist @OpenAIDevs
the “@Meta AI” app should get merged with @Muse
and the @AIatMeta One subscription needs to apply to Muse
this says a lot about American culture
this new NFL experience is actually great
that intro edit tho 😎
Another banger podcast intro from my editor 🔥
New episode dropping tomorrow with @theaaron
this is happening before the end of the month
can’t wait to get $6.32 from @XCreators then justify all my time here with:
“but babe, i get paid to tweet”
successfully ported all @grok @bot s and channels to @NousResearch Hermes, will report back on results
i love @grok @bot
i do not love its rate limits
here is a simple way to get more usage out of it:
install your coding agents on Grok Bot's virtual machine
Codex, Cursor, Grok Build, Antigravity, etc
then, install @herdrdev to manage those CLI agent sessions
now give Grok Bot a rule: use Herdr + CLI agents for anything more than conversational tasks
your rate limits for Grok Bot will only be coordination and prompting, rather than the coding tasks themselves
and you can use those @cursor_ai and Grok Build limits AND crank on Grok Bot without getting rate limited as fast
Show more
TIL you can use @NousResearch Hermes kanban cards via Telegram when on mobile
