Hi, I'm Brandon Galang, an applied-AI builder.

This is actually a very nuanced point and it's true. I say this as someone who used to do some amateur paid shoots with a Sony a7iii mirrorless

You can have a $5000 camera, but my photographer buddy with his iPhone is gonna outshoot you anyway

Similarly, anyone can code now…

pro tip if you're a gpt 6 astra maxxer, enable the experimental context management setting and adjust codex's max token limit upwards

astra is trained to use a different form of context management that leaves notes for itself. this feature is currently experimental and you need to turn it on yourself

if you're wondering how you can break into applied AI work building on the frontier, here's my simple step by step journey into applied ai:

1. using gpt 3.5, thinking it was cool but eh … 32. am now at vercel

all it took was 50B+ tokens of building and iterating

after spending 1.5 billion gpt 6 Astra tokens this weekend, here are 4 tips to level up your codex usage:

1) stay on low, crank up as you need more juice. astra low performs better than 5.6 sol high, so typically start here and work my way up to medium or high if I need codex to go deeper

from everything i've seen, gpt 6 astra really rewards those who have a strong "theory of mind"

if you cannot conceptualize the environment of your agent and the context it has access to, you are going to find gpt 6 astra disappointing

if you can, it will literally do magical things

if you're an applied AI operator, here are 3 practical Jev use cases I'm wanting to explore:

1) Jev as an 'advisor' for my agents

because Jev is so fast and cheap, you can run it alongside any agent you run to steer/stop it if it goes off track.