Zach

VMs are great but imagine giving your agent its own isolated cloud

so i have a friend. he uses a golf app at a specific driving range in SF. this app tracks his ball flight and statistics but doesnt let him export this data. so instead, he used his mac as as man-in-the-middle to capture the api host and request parameters. he learned it was a graphql server and constructed a single request that impersonated his iphone and got all 2.5k shots in one query.

@nickytonline @bcherny ideally i don’t need to install additional software

@zach_dev1 @bcherny Pomerium does this really well. usepom.link/reverse-ssh

@bcherny it would be so sick if claude code could proxy/tunnel my local services to a hostname you guys manage. i’ve built hand rolled tunneling before and im tired. please make it easy for me when i go claude rc

@MiaAI_lab @poolsideai following your git for this rn w/ dflash 🤝

Yes seems like @poolsideai updated their Laguna-S-2.1 NVFP4. Delete the old weights and redownload. I'll re-test this version.

@wicksplay but what are you building

4x DGX Spark. Another 4 in the box for right now Got some heatsinks off of Amazon that fit very nicely on the top + sides to help with heat

@WinstonLBrown1 but i know where everything is rn

@zach_dev1 Get it to clean up your desktop?

time to lock

@mike_darrow ya i should have said cook

@zach_dev1 Do you mean Cook, cause that’s what going to happen when you fire that thing up.

I take a shit picture of my dgx spark and it gets 2.2k impressions 😂

@zeeg basil hayden dark rye is solid. highly recommend

@zach_dev1 Had some finished in them but don’t recall which tbqh

It was a busy day

@CyberJeyEl let’s chat about your spark control plane

@zach_dev1 Good luck! :)

@benswerd you’re a hacker and an artist

@zach_dev1 Run the command

git.swerdlow.dev A new kind of git server. It will not let you down.

@fogle_ryan thinking about a watchdog that records everything going on in my home. computer use, voice, movement, dogs, etc. all local inference/embeddings. full control from a phone. would love to wake up to an accurate todo list that’s actively executing across machines

@zach_dev1 I tried Qwen 3.6 35B and then I tried Laguna 2.1 just did some benchmarks. What about you?

@JensenHuang i think you should follow me because im building a purpose built VMM for nvidia GPUs. 500ms provisioning, full gdr/rdma at almost line rate, for h100s/h200s/etc. qemu/cloud-hypervisor are built for everyone else (minutes provisioning time). GPUs need their own VMM

@MiaAI_lab what are you cooking

@zach_dev1 🔥🔥

@fogle_ryan what’s the first thing you ripped

@zach_dev1 Just got mine today as well!

@jonathan_conn hopefully my home lab can look similar to yours someday 🙏

@zach_dev1 Welcome to the party!