Perceptron AI ✓

@perceptroninc

The layer of intelligence that bridges the physical and digital worlds. https://t.co/Mmo3044h89

4 510Followers

0Following

94Posts total

294.7KViews on collected posts

Últimas publicaciones

Perceptron AI ✓@perceptroninc26 Aug

@perceptroninc isaac?! i’m trying this model 773 views · 7 likes · 0 reposts · 0 replies Open on X →

Perceptron AI ✓@perceptroninc26 Aug

Isaac 0.5 leads across core embodied reasoning benchmarks spanning (spatial, physical, and grounding), the dynamic nature of Isaac 0.5 delivers SOTA performance at significantly lower inference FLOPs. https://t.co/Xz4snKs3L9

2.8K views · 21 likes · 0 reposts · 1 replies Open on X →

Perceptron AI ✓@perceptroninc26 Aug

We see open-source models as essential drivers of robotics progress. That’s why Isaac 0.5 is fully open: weights, inference, and training code. With more coming soon! Blog: https://t.co/f4P4YfEqpK Weights: https://t.co/NKvQz7oT4Q Training & Inference: https://t.co/PkLYIsl1gu htt

3.1K views · 51 likes · 5 reposts · 2 replies Open on X →

Perceptron AI ✓@perceptroninc26 Aug

Cotraining perception and control allow Isaac 0.5 to make complex visual decisions during its trajectories. https://t.co/r5F8dGDhUG

2:16

2.7K views · 20 likes · 1 reposts · 1 replies Open on X →

Perceptron AI ✓@perceptroninc26 Aug

Embodied foundation models consolidate heterogeneous sources into a single model. We start with a single typed multimodal compiler (mHarmony) interleaving sources into a packed stream of tensors (TensorStream). We feed this into our latency-optimized model architecture (null http

2.6K views · 24 likes · 0 reposts · 1 replies Open on X →

Perceptron AI ✓@perceptroninc26 Aug

Scaled pretraining also enables Isaac 0.5 to convert compound commands into a series of actions. For instance, castling requires composition of “move the king to e1 to g1 and the rook from h1 to f1.” This in turn requires identifying each piece type, localizing the grid across

6:12

3.7K views · 33 likes · 1 reposts · 1 replies Open on X →

Perceptron AI ✓@perceptroninc26 Aug

A core attribute of control models continues to be elasticity: how amenable a model is to learning a new skill or task. We conduct a study measuring adaptation on a single epoch of an unseen real-world task and find Isaac 0.5 is 3x more elastic than existing open models https:/

3.2K views · 24 likes · 1 reposts · 1 replies Open on X →

Perceptron AI ✓@perceptroninc26 Aug

Today we're releasing Isaac 0.5: 36B dynamic MoE, open weight 🤗 embodied foundation model. Isaac combines multimodal video understanding, embodied reasoning and robot control into a single, sparse backbone. https://t.co/AqdVpfhESm

0:42

268.6K views · 805 likes · 104 reposts · 33 replies Open on X →

Perceptron AI ✓@perceptroninc26 Aug

We scaled video pretraining from 1,000 hours to 1 million keeping a fixed ratio of general video, egocentric, and UMI (ratio 80:30:30). This dropped teleop requirement for a well-calibrated action-loss from 5900 hrs to 28. This creates an economically efficient scaling recipe, h

7.3K views · 64 likes · 4 reposts · 2 replies Open on X →

Frente a cuentas del mismo tamaño

9 publicaciones de los últimos 90 días, junto al rango de under 10K seguidores. por debajo de sus pares en alcance y en interacción.

Visualizaciones medianas3 118esta cuenta4 175mediana de under 10K

Alcance, %69.14%esta cuenta250.82%mediana de under 10K

Interacción, %0.91%esta cuenta1.41%mediana de under 10K

Métrica	Esta cuenta	Mediana de under 10K	Proporción
Visualizaciones medianas por publicación	3 118	4 175	0.75×
Alcance (visualizaciones ÷ seguidores)	69.14%	2.5× audience	0.28×
Tasa de interacción	0.91%	1.41%	0.64×

Otras cuentas de este rango → Comparar con otra cuenta → Cómo se construyen estas referencias →

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

7.3K26 Aug

268.6K

3.2K

3.7K

2.6K

2.7K

3.1K

2.8K

773

Last 9 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

0.96%26 Aug

0.35%

0.82%

0.95%

0.96%

0.80%

1.86%

0.80%

0.91%

Reactions — likes, reposts, replies and quotes — divided by views. Median for under 10K accounts is 1.41%.

What the audience does

Likes56.5%1 049 in total

Reposts6.2%116 in total

Replies2.3%42 in total

Bookmarks35.0%650 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.

Últimas publicaciones

Frente a cuentas del mismo tamaño

Growth & engagement

Views per post

Engagement rate per post

What the audience does

Cuentas similares