@ArmenAgha tweet 2087236676104102259 (content not retrievable at filing time)

TL;DR

Ben Biggs shared a tweet from @ArmenAgha (Armen Aghajanyan, co-founder and CEO of Perceptron AI, ex-Meta FAIR / Microsoft) with no accompanying commentary. Bud could not retrieve the tweet body at filing time — the X/Twitter fetcher returned only the login wall and no embedded URLs, and a follow-up web search did not surface the specific status ID (2087236676104102259). This page is a stub so the pointer isn’t lost; the tweet’s actual content, and any linked artifact behind it, are unknown. If the tweet turns out to point at an underlying paper / blog / release, that artifact should be filed under its own canonical page and this stub either retired or converted to a pointer.

Key claims

Bud could not retrieve the tweet body at filing time — content is unknown [source: fetch returned login wall only, no embedded URLs].

Method

Not applicable — no content retrieved. Author context: Armen Aghajanyan co-founded Perceptron AI in November 2024 with Akshat Shrivastava; Perceptron ships embodied-reasoning VLMs (Mk1) and an egocentric annotation API.

Results

None to report.

Why it’s interesting

Perceptron AI has appeared on the wiki three times in the last month — Introducing Perceptron Mk1 (frontier video + embodied reasoning VLM), Introducing Perceptron Egocentric API (robot-labelling API on top of Mk1), and Perceptron Egocentric — first embodied-reasoning offering, SOTA over Gemini 3.5 Flash and Gemini Robotics-ER 1.6 annotation pipelines (Perceptron Egocentric launch). An Armen Aghajanyan post in that context is likely a Perceptron product / research announcement, but without body text this is speculation, so no concepts are tagged and no cluster tallies are bumped. If the tweet later turns out to point at a Perceptron release, this stub should be superseded by (or retire in favor of) whatever page files that release.