Elliot Arledge ✓

@elliotarledge · 4 x RTX PRO 6000 · joined 25 Nov 2022

@infinity_ai_ https://t.co/82V5vFob7t | made the 12 hr CUDA course | https://t.co/2hzezpo5Hi | @shipfr8 alum

43 369Followers

345Following

16 626Posts total

1.8MViews on collected posts

Derniers posts

Elliot Arledge ✓@elliotarledge4 Sep

gemini 3.8: Bottleneck is the 600-watt board power ceiling. 126 views · 0 likes · 0 reposts · 0 replies Open on X →

Elliot Arledge ✓@elliotarledge4 Sep

Astra cooking delegating groks for execution while I'm on starlink wifi (coming back to canada! more timelapses coming) https://t.co/RVNar6Zrx2 496 views · 4 likes · 1 reposts · 0 replies Open on X →

Elliot Arledge ✓@elliotarledge4 Sep

astra reading gemini 3.8s slop https://t.co/3Lv1tlnMWG 513 views · 2 likes · 0 reposts · 0 replies Open on X →

Elliot Arledge ✓@elliotarledge4 Sep

the strat of buying all the coding plans across all providers (maybe 3 to 7 of them) and having a solid "delegate" skill for spreading token usage evenly is the more economical way to move faster. 761 views · 10 likes · 0 reposts · 0 replies Open on X →

Elliot Arledge ✓@elliotarledge29 Jul

@elliotarledge https://t.co/gXWQSNNYAg 12.3K views · 116 likes · 1 reposts · 3 replies Open on X →

Elliot Arledge ✓@elliotarledge29 Jul

Numbers (RTX PRO 6000 / 9950X3D, methods + harnesses in the repo): > exact render, 1080p: 35.9 fps GPU, 4.5 CPU > batched sim: 3.0M env-ticks/s at N=8192, 756k agent steps/s The CPU one is bit-identical. The slow one is the oracle. 40.8K views · 135 likes · 1 reposts · 1 replies Open on X →

Elliot Arledge ✓@elliotarledge29 Jul

C, CUDA, no engine, no game jar at runtime. You bring your own copy of the game; every Mojang-derived input regenerates locally, byte-identical. I'm excited to see any and all constructive criticism and what the community can do with this. kernels + source + oracle mod: 40.1K views · 384 likes · 3 reposts · 10 replies Open on X →

Elliot Arledge ✓@elliotarledge29 Jul

Training mode throws the full rasterizer away. The policy sees a 64x36 block-id camera with depth and edge planes - right pane is the whole observation. That is why thousands of worlds fit on one card. https://t.co/x29AxnQPoB 44.4K views · 227 likes · 0 reposts · 1 replies Open on X →

Elliot Arledge ✓@elliotarledge29 Jul

2,058 actions, one unbroken chain, nothing scripted: spawn empty-handed -> logs -> planks -> sticks -> crafting table -> wooden pickaxe -> stone -> coal. This is the middle of it, unedited, real speed. https://t.co/mx3SGf6phh 50.7K views · 157 likes · 0 reposts · 2 replies Open on X →

Elliot Arledge ✓@elliotarledge29 Jul

Why exactness: it makes an RL sim you can trust. 23/23 recorded tapes replay with physics clean to 1e-9, CPU and CUDA bitwise-identical every tick (all the way to defeating the ender dragon). Train here, replay the same actions in the real Java game, get the same result. There 52.3K views · 245 likes · 0 reposts · 2 replies Open on X →

Elliot Arledge ✓@elliotarledge29 Jul

Left is Mojang's game. Right is mine, replaying the same inputs. Validated on terrain, physics, and pixels - the gates ship in the repo. Can you spot the difference? https://t.co/LeLwt572VT 85.9K views · 429 likes · 4 reposts · 5 replies Open on X →

Elliot Arledge ✓@elliotarledge29 Jul

The receipts: my inventory screen minus Minecraft's, amplified 80x so you can see anything at all. 112,796 pixels. 442 differ, each by 0.3% https://t.co/AjbaMrdO2h 62.2K views · 324 likes · 1 reposts · 2 replies Open on X →

Elliot Arledge ✓@elliotarledge29 Jul

Introducing Netherite! Minecraft 1.11.2, rewritten from scratch in C and CUDA, bit-verified against the real game. This is one trained agent playing in it - then the renderer it actually trains through - then 7,200 live worlds stepping in lockstep on one GPU. https://t.co/eg2gX 1.4M views · 4.8K likes · 197 reposts · 160 replies Open on X →

Face aux comptes de taille comparable

4 posts des 90 derniers jours, à côté de la tranche de 10K–100K abonnés. en dessous de ses pairs à la fois en portée et en engagement.

Vues médianes504ce compte924médiane pour 10K–100K

Portée, %1.16%ce compte3.62%médiane pour 10K–100K

Engagement, %0.70%ce compte1.52%médiane pour 10K–100K

Indicateur	Ce compte	Médiane pour 10K–100K	Rapport
Vues médianes par post	504	924	0.55×
Portée (vues ÷ abonnés)	1.16%	3.62%	0.32×
Taux d'engagement	0.70%	1.52%	0.46×

Autres comptes de cette tranche → Comparer avec un autre compte → Comment ces repères sont établis →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.

Views per post

1.4M29 Jul

62.2K

85.9K

52.3K

50.7K

44.4K

40.1K

40.8K

12.3K

7614 Sep

513

496

126

Last 13 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

0.37%29 Jul

0.53%

0.51%

0.48%

0.31%

0.52%

0.99%

0.34%

1.00%

1.31%4 Sep

0.39%

1.01%

0.00%

Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.52%.

What the audience does

Likes69.2%6 815 in total

Reposts2.1%208 in total

Replies1.9%186 in total

Quotes0.8%76 in total

Bookmarks26.1%2 567 in total

Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.

The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.

Derniers posts

Face aux comptes de taille comparable

Growth & engagement

Views per post

Engagement rate per post

What the audience does

Comptes similaires