Spencer Cooley

Aerospace Engineer. Providing commentary as we traverse the rapidly changing world.

[Spencer Cooley @srcooley3]

https://t.co/6D4bsORd5s

[Spencer Cooley @srcooley3]

@EdwardAlerts $100

[Edward Johnson @EdwardAlerts]

I AM OFFICIALLY RESTARTING THE $100 TO $100,000 CHALLENGE NEXT MONDAY FOR COMPLETELY FREE☢️

THIS WILL BE DONE IN A SECRET X GROUP CHAT WHERE I WILL POST ALL MY TRADES WITH ENTRY AND EXIT FOR FREE🏄‍♂️

LIKE, REPOST, & REPLY “$100” TO SECURE YOUR SPOT🔥

YOU MUST BE FOLLOWING ME💰

[Spencer Cooley @srcooley3]

@vulturetrades Do you think hold or cut losses if we’re still in? Current sitting at 0.23 and I put a sell order for 0.25 but would prefer to get back in the green if you think it may turn around

[Spencer Cooley @srcooley3]

@EdwardAlerts PDF

[Spencer Cooley @srcooley3]

@EdwardAlerts Trade

[Spencer Cooley @srcooley3]

@rileybrown @rileybrown haha shit just saw your earlier post, guess we’re on the same wavelength, excited to see what you’re working on!

I was close to buying a cheap android phone and hacking it to just run an openclaw instance and nothing else so it was an AI edge node

[Spencer Cooley @srcooley3]

@rileybrown I’m almost thinking we bypass the “app” paradigm completely just haven’t teased out exactly what will replace it. Your browser video I think was a great hint at that. All apps do is log data, present data, and take action (games & social media are exceptions)

[Spencer Cooley @srcooley3]

@rileybrown There will be no more apps, all devices will become edge nodes and interfaces for AI and will generate data collection/visualization in whatever format you desire.

The only “apps” that will remain will be social media because you need a centralized connection for functionality

[Spencer Cooley @srcooley3]

@elonmusk @DrewPavlou Are floating launch sites also a consideration?

[Spencer Cooley @srcooley3]

@jun_song Extreme budget setup here is Mac mini and iPad with workbench, my total cost was $550 + $20/mo (well plus $100/mo for codex subscription)

[Spencer Cooley @srcooley3]

I believe allowing Anthropic to use Colossus 1 wasn’t a sign that grok was dead, it was a sign that they were so confident in where grok development was headed they could afford to let their competitor use their hardware.

The secret weapon is the vision training data from Tesla.

Digital Optimus, or what could be called Grok Computer Use, is going to be the product that comes out this year and completely destroys everyone else.

Utilizing their vision training data appropriately to view the computer screen and control the mouse and keyboard will be orders of magnitude better than what Perplexity, Claude, or Codex currently do.

This computer use is the holy grail of Agentic AI, it bypasses all the tricky integrations and connections that define the current agent space.

I just realized… perhaps Cursor is actually the perfect name for this

[Spencer Cooley @srcooley3]

@kapilansh_twt Yes, just have backups

[Spencer Cooley @srcooley3]

@TheGeorgePu They’re just selling out before switching to the new model

[Spencer Cooley @srcooley3]

Why would they do that? They have 4 models in training right now and will have first access to all the compute, and then for Tesla they have the Optimus robot and digital Optimus which they will want to run on their own model.

I would only agree with that if Elon somehow has complete faith that Anthropic or some company is honest in their mission statement and it aligns with groks current mission statement.

I think they’ll continue to lag for a year or so but they’ll overtake very quickly once everything’s online

[Spencer Cooley @srcooley3]

@NoLimitGains Meaning what? Time to sell everything?

[Spencer Cooley @srcooley3]

@strengthPlan Hahaha this is exactly how I’m feeling right now, gotta remind myself I’m on a long term horizon, but man the impatience is real

[Spencer Cooley @srcooley3]

@OpenAI I wonder if this will be the best model to use with openclaw for day to day conversations

[Spencer Cooley @srcooley3]

@BenjaminBadejo I’m having issues with codex harness, what advantages have you found? Pi seems much better for daily communication and the agent can still spawn a codex session when needed

[Spencer Cooley @srcooley3]

I’m having issues with using codex runtime vs native pi runtime with gpt. Have you experienced any issues? The codex runtime feels significantly slower, also I use a different model for my cron jobs and they’re failing because codex searches for those models and doesn’t find them so they fail.

For now I switched back to pi runtime, but I’ll update to 2026.5.2 to see if there’s any improvement

[Spencer Cooley @srcooley3]

@elvissun They didn’t raise prices, the 256gb is sold out