Ejaaz (@cryptopunk7213)

Pinned post

Flight delayed 7 hours - asked Muse to file for compensation. 5 mins later i had $250 credit in my delta account. it even found and rebooked me a new flight.

it just figured everything out. even responded to the support email itself.

shit feels like magic.

Ejaaz @cryptopunk7213 3h

dude on vacation decided to ask AI to discover hidden ancient civilizations in Mexico.

72 hours later, 48 ai agents found an unmapped Maya site using publicly available NASA LiDAR data and the Mexican government confirmed it.

this man had 0% experience in archaeology. you can just go out and do things now.

Pietro Miotti Verified account @PietroMiotti 1d

My AI agents found an unmapped Maya site in Mexico, using LiDAR data from a NASA dataset.

72 hours from a random thought on vacation to a reply from Mexico's national archaeology institute.

What a time to be alive…

View quoted post on X

Ejaaz @cryptopunk7213 5h

demand for anthropic’s IPO is off the charts, the entire $64 billion raise is so oversubscribed they’ll likely raise over $100B next week

this would be a record raise beating the spaceX IPO.

fable 5.2, opus 5.5 are great models but investors are pricing in what anthropic is hiding behind the scenes - internal models that are accelerating AI R&D 2X per year

those models will bring in incredible revenue for the company.

Ejaaz Verified account @cryptopunk7213 1d

so the anthropic IPO valuation will be around the $1-1.3 trillion range, much healthier than the original 2T

with a $80 billion ARR (as of oct 31st) plus $200B+ in cash i think this puts anthropic in a great position. revenue is growing at an insane rate and demand will now increase with fable 5.2 launch.

worth noting also that 7 of anthropic's biggest suppliers are being given a private placement of $26 billion in the IPO. more of an alignment stake tbh but worth keeping an eye on.
View quoted post on X

Ejaaz @cryptopunk7213 15h

do not for a second think microsoft and google aren't worried about this. this is chipping away at their business

these AI office tools will be served to billions of claude/GPT users giving them yet another reason to stay in those ecosystems

theres a reason chatgpt now makes up 20% of google's search volume.

think of AI as a magnet, the more it can serve your existing needs the stickier the lab's moats become.

who cares if the 1st versions of these tools suck? they'll improve quicker and surpass existing tools in weeks at the rate models are improving.

the cost of software is officially zero. staying power is being chipped away by every new claude/gpt release.

Ejaaz Verified account @cryptopunk7213 1d

lol anthropic just automated 90% of every product manager, data engineering and finance team job.

now a single person can do all of this and have AI review and steer the whole process:

- multiple brainstorming sessions
- hours and hours of tedious back and forth, creating slide decks, writing memos.
- presentations to board and managers.

remarkable.
View quoted post on X

Ejaaz @cryptopunk7213 16h

woah if Apple really is building a way to let you run open source models locally on your macbooks without leaking data then this is will turn them into the #1 AI hardware network worth billions.

so many companies are freaking out about anthropic and openai copy-pasting their entire product, the solution is privately hosted models but they're expensive to setup and run.

Apple's hardware could be the best compromise. connect a network of macbooks, mac mini pro's and inference the shit out any model.

Aaron Verified account @aaronp613 1d

Apple is working on a new app called Harbor

Harbor turns your Mac into a private AI inference server. Run open-weight models locally on Apple silicon—your data never leaves your hardware, and nothing reaches the cloud. Connect multiple Mac computers over Thunderbolt 5 to pool unified memory, enabling large models and higher concurrency that a single computer can handle alone.

Set up in minutes, not hours Harbor automatically detects your devices over Thunderbolt 5 to form a cluster. Plug in, start, and you’re running—no manual configuration, no command line, no infrastructure expertise required.

Download and run models directly Browse, download, and load open-weight models from within the app. Harbor handles model management so you can focus on the work, not the setup.

Cluster for scale Connect two or more Thunderbolt 5-capable Mac computers and Harbor pools their unified memory into a single inference resource. Run models that exceed any single computer’s capacity, or serve more concurrent requests across your cluster.

Harbor is currently in a private preview under the Claris brand.

View quoted post on X

Ejaaz

Latest posts 5