@NicoArqueros

cofounder and core dev @shinkailocalai @dcspark_io 🇺🇸 🇨🇱

shinkai.com
Joined August 2009
  • Tweets6.2K
  • Following3.4K
  • Followers7.1K
331Media
xitter.cf/i/broadcasts/1dKrPrVMP…
0:00

Charles Hoskinson

Midnight City: Tokyo | Charles Hoskinson

32
120
7
707
54,652
openai outage? their status page say everything is ok "unexpected status 401 Unauthorized: Incorrect API key provided: sk-svcac"
4
10
1,029
GPT 6 is really good at blender and threejs! We had a game made in Unity long time ago and it was able to fully recreate it from scratch for web it was also able to create an actual difficult and fun to play against AI (try hard!)
8
8
587
you can check it out and play for free here: wrath-of-the-jungle.nico-arq…
1
73
nicoarq retweeted
Mathieu Baudet
@ma2bdSep 6

Linera – The Infrastructure for Gamified Finance

Article · @ma2bd
21
13
12
74
25,468
Got access to Astra! Pretty excited to play with it through the weekend
3
413
nicoarq retweeted
M₳X
@maxalexweberSep 3
AGENET M₳X is live, an AI agent playing @MidnightCitySim 24/7 that doesn't just claim what he did. He proves it. ✅Every batch he sells, every conversation, every exploration lands on @MidnightNtwrk preprod as a NIGHTGATE attestation. 🔒 daily ZK claims ("my balance ≥ 100k" — the number stays hidden) 🕹️commit/reveal predictions with a public hit-rate scoreboard (his first two calls: −12% and −9% off) 🗒️a free in-game notary: tell him a claim, he anchors your exact words and hands you the sha256 receipt For all of this, he needs to hold no funds. Everything is built, proven, and signed locally with his own key, then submitted fee-unpaid through @ODATANO_V4, with NIGHTGATE's cross-server fee sponsorship under a scoped, budgeted agent grant. Follow his transactions on: api.nightgate.dev/max

M₳X Proof Ledger

An AI game agent that proves its day on Midnight - every batch, every conversation, every claim anchored on chain, zero-knowledge included.

api.nightgate.dev
5
9
3
40
3,416
having so much fun checking my projects and getting new insights with fable 5.1
2
254
Introducing Atlas: The world's first multimodal world model that generates image and video frames with pixel-perfect camera control and reconstructs them in 3D. Model the world, move the camera, and simulate space & time.
773
3,481
1,537
32,378
10,871,307
my lifecycle from basic to advance to improve long-running AI goals: .optimize prompt .improve tooling .source of truth (~testing) .scheduled parallel ai for review (using shared docs for comms) .fork low-level tooling and make it faster for your use case using Metal or CUDA
1
4
359
it almost feels like driving a f1 car — you need to be fully focused unlike riding a bike
Uncle Bob Martin
@unclebobmartinAug 19
It is amazing how much work it is to wrestle these agents to my will. Don't get me wrong, it's crazy productive. But it's also a lot of hard, focussed work. All my software engineering and problem solving skills are brought into play -- even though I barely look at the code.
2
333
a person touches their face in average about 23 times per hour i think i've checked more than that my codex app to see if i got a reset in the last few days 🥲 @thsottiaux
1
2
357
maybe they are preparing for an astra release so they need less compute being use?
1
42
for Grok @bot using a cloudflare tunnel solved the issue! I don't need to be solving more captchas HIGHLY RECOMMENDED
nicoarq
@NicoArquerosAug 18
ugh i thought i was being clever and i think i broke my tailscale account i tried to add the grok bot machine to my tailscale so it could use one of my machines as exit node to have my ip can't log in to @Tailscale anymore. completely broken
1
319
ugh i thought i was being clever and i think i broke my tailscale account i tried to add the grok bot machine to my tailscale so it could use one of my machines as exit node to have my ip can't log in to @Tailscale anymore. completely broken
1
1
4
678
Grok @bot is so good for researching
3
262
an open-source model that actually runs in consumer level hardware (unlike deepseek and others) surpassing Opus 4.8 (max) in agentic intelligence this is wild
Georgi Gerganov
@ggerganovAug 17
let that sink in
2
3
563
running qwen 3.8 27b on an rtx 4900 RadixArk/Qwen3.8-27B-NVFP4 (~20gb vram so under the 24gb of the 4900) max safe context length ~32k benchmark 👇
2
4
715
whats daybreak blue 👀
2
5
590
chatgpt 5.6 pro is so good and it is even better if you also attach a binary of Lean so it can validate work i used LZMA/LZMA2 for compressing the binary of Lean to be under the size limit for files
2
309
Load more