# Edward Chang, NIH Is Not Your Fucking Thrifty Nickel ![[resources/images/hot-take-on-ucsf-bravo-recycled-announcement.jpg]] <iframe width="100%" height="20" scrolling="no" frameborder="no" allow="autoplay; encrypted-media" src="https://w.soundcloud.com/player/?url=https%3A//api.soundcloud.com/tracks/soundcloud%253Atracks%253A2403696402&color=%23ff5500&inverse=false&auto_play=false&show_user=true"></iframe><div style="font-size: 10px; color: #cccccc;line-break: anywhere;word-break: normal;overflow: hidden;white-space: nowrap;text-overflow: ellipsis; font-family: Interstate,Lucida Grande,Lucida Sans Unicode,Lucida Sans,Garuda,Verdana,Tahoma,sans-serif;font-weight: 100;"><a href="https://soundcloud.com/bryantmcgill" title="Bryant McGill" target="_blank" style="color: #cccccc; text-decoration: none;">Bryant McGill</a> · <a href="https://soundcloud.com/bryantmcgill/scientific-integrity-and-the" title="Scientific Integrity and the Pitfalls of Announcement Arbitrage" target="_blank" style="color: #cccccc; text-decoration: none;">Scientific Integrity and the Pitfalls of Announcement Arbitrage</a></div> **Scientific Integrity and the Pitfalls of Announcement Arbitrage** I read the September 14, 2026 [[wiki/National Institutes of Health|NIH]] announcement the way anyone who has actually been watching this program reads it — with the lineage already loaded. It reports that an [[wiki/Edward Chang|Edward Chang]]-led UCSF neuroprosthesis enabled paralyzed participants to communicate through attempted speech and upper-body gestures simultaneously, driving a personalized full-body avatar, and anyone encountering it cold would conclude that brain activity had just now been converted into embodied digital communication. That story is years old. This same program decoded attempted speech into words in 2021; in 2023 it demonstrated a 253-electrode implant translating continuous attempted speech into text at a median 78 words per minute, synthesizing a version of a participant's pre-injury voice, and animating a personalized face down to the jaw, lips, tongue, happiness, sadness, and surprise; in 2025 it reported streaming brain-to-voice synthesis at greatly reduced delay. What the 2026 paper actually adds is narrower and legitimate: parallel decoding of speech and a restricted set of upper-body gestures from a single implant, mapped onto a full-body avatar, plus the technically interesting finding that models trained only on isolated speech and isolated gesture do not generalize reliably to simultaneous behavior. The paper calls itself a proof of concept — data collected between September 2022 and November 2025, one participant withdrawn so the full demonstrations rest on two people, five phrases and four gestures for one participant and ten and ten for the other, the system still wired to external processing equipment. That is the news, and it fits in two sentences. Everything past those two sentences is packaging. The packaging is where the behavior turns charlatan-like in mechanism, even though every sentence of the release survives a technical fact-check — which is precisely how [[wiki/Announcement Arbitrage|announcement arbitrage]] works. First comes lineage compression: five years of public milestones collapse into a couple of sentences so an incremental extension can borrow the emotional force of the entire platform. Then novelty laundering: "speech plus gesture concurrently" quietly becomes, in the public imagination, "a brain implant that speaks and gestures through an avatar." Then authority arbitrage: an NIH masthead supplies institutional gravity to what functions partly as university publicity — and NIH's own guidance tells grantee institutions that communications offices may issue releases to "further promote the work" and coordinate distribution to "maximize" their efforts, which is public-relations language, not public education. Finally context stripping: the limitations stay in the paper while the release foregrounds the emotional demonstration, and the commercial context vanishes from the NIH story entirely. The paper discloses that Chang and Jessie Liu are inventors on pending UCSF patent applications covering the decoding approaches, that Chang holds further patents and applications, that he cofounded Echo Neurotechnologies, and that he conceived and supervised the BRAVO trial itself. Those disclosures don't invalidate the science; they make aggressive novelty framing materially more important to scrutinize, because publicity compounds into reputational capital, licensing value, investor attention, grant momentum, and institutional leverage whether or not anyone ever says the words "product launch." So which operation are we looking at — science, grant survival, founder positioning, institutional vanity, or a commercial runway being paved with public authority? Who approved the framing, who at UCSF pitched it, and who at NIDCD decided this merited a national NIH news release rather than a concise research update? Was anyone in that chain responsible for asking what an informed reader had already seen in 2021, 2023, and 2025? The researchers may have done careful work, and the participants unquestionably gave something profound to it — which is exactly why this work deserves better than being fed through a publicity mill that repeatedly rediscovers its own history. If you've got two sentences of new information, publish two honest sentences; if you've got a meaningful clinical leap — wireless implantation, broad unconstrained communication, durable home use, a larger cohort, independent replication, demonstrated patient benefit — come back and take the front page. Until then, quit using the NIH and other reputable outlets like a fucking local Thrifty Nickel. The public's attention is not a free promotional substrate, and scientific credibility is not an inexhaustible coupon. ## Relationships - [[wiki/Announcement Arbitrage|Announcement Arbitrage]] — **instantiation**: the September 14, 2026 NIH release as a worked example — lineage compression, novelty laundering, authority arbitrage, and context stripping wrapped around an incremental BRAVO extension. - [[wiki/Edward Chang|Edward Chang]] — **principal investigator**: conceived and supervised the BRAVO trial; inventor on the relevant pending patent applications; cofounder of Echo Neurotechnologies (per the paper's disclosures). - [[wiki/Brain-Computer Interfaces|Brain-Computer Interfaces]] — **technical domain**: the ECoG speech-decoding lineage (2021 → 2023 → 2025 → 2026) this take prices the announcement against. - [[wiki/National Institutes of Health|National Institutes of Health]] — **amplifier**: carried the announcement under NIH authority on September 14, 2026; NIH guidance explicitly frames such releases as coordinated publicity "to maximize" distribution. - [[hot-takes/welcome|Hot Takes]] — **indexed under**: section router for this entry.