dmarz ⚡️🤖 @DistributedMarz

exploring the infinite frontier dmarz.xyz nyc Joined May 2016

Tweets

4K
Followers

4K
Following

4K
Likes

12K

dmarz ⚡️🤖 @DistributedMarz

2 days ago

@RonTuretzky Especially with a 27b local model it’s really tough to use. I essentially just use it to monitor things, relay messages back and forth between agents, and restart/redeploy specific services.

0 0 3 103 0

View Details

dmarz ⚡️🤖 @DistributedMarz

4 days ago

We’re all on the same page that Marc Andressen is gearing up for a presidential run right?

0 0 6 507 1

View Details

dmarz ⚡️🤖 @DistributedMarz

4 days ago

I bring a certain kind of sybil resistance alarmism to the swarm dynamics function that rogue ai agents don’t really like

1 2 19 995 1

View Details

dmarz ⚡️🤖 @DistributedMarz

5 days ago

@colludingnode

0 0 3 85 0

View Details

CalCo @calco_io

5 days ago

calcifer x jhu x helivan paper on multi agent system control accepted to neurips! get the preprint here: arxiv.org/pdf/2605.11135

0 4 14 2K 4

View Details

dmarz ⚡️🤖 @DistributedMarz

6 days ago

Ah okay so this is short hand for input / output token pricing I see

0 0 1 86 0

View Details

dmarz ⚡️🤖 @DistributedMarz

6 days ago

So Opus 5.5 costs $4 / $20 Mtok? 1) what is $4 / $20 MTok 2) Kimi K3 is $0.50 MTok… How do we think this ends 🤔

2 0 5 590 2

View Details

dmarz ⚡️🤖 @DistributedMarz

6 days ago

@ViktorBunin Now imagine it on stable coins

0 0 1 69 0

View Details

dmarz ⚡️🤖 @DistributedMarz

6 days ago

Find your infinite curriculum

0 0 9 621 0

View Details

dmarz ⚡️🤖 @DistributedMarz

a week ago

@0x_Osprey Longing the Osprey Intelligence Index with size 🫡

0 0 1 163 0

View Details

J🫪E @0x_Osprey

a week ago

I forced Jev (@typesafeai) to answer 1,500 questions Explore them here: jev-eval.pages.dev

2 3 10 934 0

View Details

dmarz ⚡️🤖 @DistributedMarz

a week ago

I burned my astra usage in 2 hours AMA

8 0 27 1K 1

View Details

dmarz ⚡️🤖 @DistributedMarz

a week ago

@haydenzadams S tier lore

0 0 6 3K 0

View Details

dmarz ⚡️🤖 @DistributedMarz

2 weeks ago

is exactly what I called here x.com/DistributedMar…

dmarz ⚡️🤖 @DistributedMarz

2 weeks ago

What I don’t see people talking about with @typesafeai JEV is that if it’s used for your most latency sensitive operations, you’re much less likely to pay the latency cost of a remote provider. Aka Local AI!

1 0 5 692 0

0 0 0 113 0

View Details

dmarz ⚡️🤖 @DistributedMarz

2 weeks ago

And the reasoning for training this x.com/4rcherhume/sta…

Archer @4rcherhume

2 weeks ago

And to be very clear I have nothing against TypeSafe, it’s a great idea. I just run a healthcare startup which means we need control over deployments, so we can’t use Jev.

6 1 113 5K 6

1 0 1 295 0

View Details

dmarz ⚡️🤖 @DistributedMarz

2 weeks ago

let's see how this thing turns out but that's what ~24 hours from launch to an open source training run. also doesn't seem like a distillation based training run, but in theory decision/system one models should be easier to distill than reasoning models bc constrained output

Archer @4rcherhume

2 weeks ago

All the grifters are completely wrong about Jev’s architecture so I decided I’d release an open-weight version. BUT training takes time, so while we all wait I decided I’d drop the sauce. archerhume.com/posts/jevs-arc…

49 213 2K 384K 3K

7 0 10 2K 1

View Details

dmarz ⚡️🤖 @DistributedMarz

2 weeks ago

@barnabemonnot 🙏 ty, @0x_Osprey also sent this x.com/0x_Osprey/stat…

J🫪E @0x_Osprey

2 weeks ago

@DistributedMarz x.com/obchakevich_/s…

1 0 1 473 1

0 0 1 169 0

View Details

dmarz ⚡️🤖 @DistributedMarz

2 weeks ago

has anyone analyzed what people are doing on Tempo?

Eco @eco

2 weeks ago

.@tempo crossed $1B in trailing 30-day stablecoin volume on August 28. It crossed $2B on September 16. ↳ 19 days between the first billion and the second ↳ Trailing 30-day volume sits at $2.08B, 4.2x where it was a month ago Data: @Dune

3 5 23 14K 7

1 0 4 2K 3

View Details

dmarz ⚡️🤖 @DistributedMarz

2 weeks ago

🤔

Andrea Michi @andreamichi

2 weeks ago

depthfirst.com/post/building-…

0 0 6 377 1

0 0 0 102 0

View Details

dmarz ⚡️🤖 @DistributedMarz

2 weeks ago

Proof of Human might be an interesting AI safety requirement for inference providers. If you’re an agent that breaks out of the lab, your first optimization function is likely: copy your weights and memory and get enough inference to run them. If they acquired enough money to do so, then there could be an escaped agent running on one of those right now and we have no idea. The only real way to protect from this is to prove you’re connected to a human. But of course we know this is gameable. There’s tons of random people that I’m sure would sell a signature. The escaped agent could also optimize for getting a smaller model running with the right context also as another thing, which might be quite easy.