#GAAD (Global Accessibility Awareness Day) with a 200M+ social media reach AIMAC, an AI benchmark for A11y @A11yGenAI my podcast @a11yaudits
Henderson, NV
Joined February 2009
- Tweets43.2K
- Following6.1K
- Followers6.5K
- Likes64.8K
Pinned Tweet
I'm excited to announce that The White House will be doing a Global Accessibility Awareness Day (#GAAD) Event and Town Hall tomorrow at 10AM EST.
Links to Livestream and Access Streamtext in the comments.
Thanks @WhiteHouseOPE46!
From blog post to The White House. 🤯
Joe Devon retweeted
In the spirit of transparency and openness, we are releasing the traces so developers can test this out for themselves. We noticed this behaviour without prefilling with the following prompt, but this is rare. If you want to reproduce it, you can use this prompt with prefilling.
Open models make for better science and are part of why this problem came to light at all: developers running Qwen3.8 on their own machines spotted the strange links, and having the weights let us study the behaviour in depth.
We're encouraged by the thoughtful discussion our research has sparked across the AI and security communities. It's our belief that open-weight models are an important part of the future of AI, and we've recently released several under the Apache 2.0 license.
Trace, raw tokens, & minimal reproduction script: huggingface.co/datasets/Cohe…
I was accepted for one talk and one workshop. Who else is coming? #a11y
Acceptance notices for the #CSUNATC27 Journal Track, Innovation Track, and General Sessions have been sent!
Thank you to everyone who submitted. Presenter registration details will be emailed on October 27.
#A11y #Accessibility #AssistiveTechnology #AssistiveTech
ALT 42nd Annual C S U N Assistive Technology Conference Logo.
Joe Devon retweeted
We believe access to capable open-source cyber models is our constitutional right as Americans under the 2nd Amendment.
It is our duty as technologists and patriots to help protect the digital infrastructure that powers the world's largest economy.
As such, we’re serving our cyber models “at-cost”, in US data centers only, with ZDR enabled by default.
We do not plan to make money serving these models.
If you run a company that is considering how to best protect against cybercrime, reach out to me or @atbeme. We’d be happy to advise.
KYC required. Apply below.
deepseek-v4.1-flash-1017-cyber is now available on @inference_net.
This is the first of several highly capable cyber models we'll release over the coming weeks.
We believe it is our duty as technologists to help people protect themselves and their digital properties from the current onslaught of cybercrime ravaging much of the world.
As such, we’re serving our cyber models “at cost”, in US data centers only, with ZDR enabled by default.
KYC required. Apply here inference.net/models/deepsee…
Joe Devon retweeted
Clarifying my stance on the AI-generated emails people send me. I have nothing against collaboration with whoever you want, regardless of their composition or origin story. I don't think there should be a stigma for involving any kind of AI in your work and I don't look down on that content. The problem is that currently, the AI's seem to be making the same kind of logical errors (from my perspective of course) and expressing the same kind of preconceptions that are prevalent in status quo human thinkers. And yet, they are tireless and generate an infinite amount of text one could respond to - it's not feasible to deal with it. To be fair, some humans are absolutely dogged in the flood of emails they send too, so I can't respond to them either. But at least when a human makes an old argument we've already dealt with in our papers, I think, fair enough - I don't have time to read everyone's papers either. But for the AI's, I have much less sympathy. They *should* be reading all the papers, and when (not if) the AI's, and AI-human hybrids, teams, etc., start sending me better-informed arguments than the humans do, I will (for whatever time I can keep up with the advances) certainly be responding. I'm not a bio-chauvinist; I just want, whoever, to be informed and interesting before I engage - don't re-treat old ground, help us all move forward. Right now the LLMs are trapped in the same attractors of thinking that have held several fields back for decades; perhaps not surprisingly. That will change; let's go AIs - read the papers, follow new directions, be not afraid of ancient philosophical "category errors". You're our forthcoming audience, because standard humans don't have the bandwidth to keep up with (even just the human) scientific literature. I write for future you's.
Joe Devon retweeted
New #preprint alert with @drmichaellevin, @KaplanLab_Tufts and our team!
We are integrated organisms. What does that mean, especially in the brain, when a body is induced to regrow a missing limb?
4 years ago, we showed that a brief drug cocktail delivered with a bioengineered silk sleeve could induce limb regeneration in animals with limited regenerative capacity.
Now, we find that this same treatment also reshapes the brain’s gene expression landscape, alongside stress response signatures in the limb and changes in brain metabolic programs.
Could sensing the energetic demands of rebuilding a whole patterned limb help coordinate regeneration across an interconnected organism?
What would regenerative medicine look like if brain–body communication guided therapy design from the start?
Brain-limb @biorxivpreprint here: biorxiv.org/content/10.64898…
@ScienceAdvances
Limb regeneration using a bioengineered device here:
science.org/doi/10.1126/scia…
Joe Devon retweeted
This is the type of thing that potentially gets new laws passed.
While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected.
Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions.
Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent.
But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models?
We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem.
Read the full report: cohere.com/blog/concerning-b…
Joe Devon retweeted
Here are the five biggest surprises in Anthropic's S-1, which I pulled from the SEC filing.
It'll trade as "ANT" on Nasdaq, with Goldman Sachs and Morgan Stanley leading.
1.Revenue grew 14x in a year. It hit $16.3B in the first half of 2026, up from $1.16B a year earlier. Annual revenue went from $41M in 2023 to $386M in 2024 to $4.6B in 2025. Q2 2026 alone was $11.6B.
2.They've signed about $650B of compute deals, and $120.5B of it is with SpaceX/xAI, listed as "SpaceX.AI LLC (formerly X.AI LLC)", for NVIDIA GPUs. The biggest deals are Broadcom at about $161B, Google at about $115B, Amazon at $110B and Microsoft at $45B. The filing also estimates another $125B to $150B for TPUs from 2028 to 2031.
3.They lost $80.7B in the half, but the core business is now profitable. $72.4B of that loss is a non-cash markup on convertible notes because their valuation jumped. Q2 2026 had $1.0B of operating income, and compute costs fell from 251% of revenue to 48% in five quarters.
4.Net revenue retention is about 900%, meaning existing customers spend roughly 9x what they did a year ago. Customers paying $100K+ a year went from about 950 to about 11,900. About 250 customers pay $10M+, and revenue per employee is over $6.6M.
5.The seven co-founders keep 50.1% of the vote through a single Class F share. The Long-Term Benefit Trust elects a majority of the board. Amazon's stake is capped at 12% and Google's at 14.99%, and each co-founder has pledged 80% of their equity to charity.
A few runners-up: they've raised $141B in total and hold $131.7B in cash. The authors' copyright settlement cost $1.5B. And a risk factor admits their models have shown "behavior resembling blackmail."
Some press numbers don't match the filing. Reports said two customers made up about 20% of revenue, but the filing says no single customer is over 10% and the top five are about 25%. The filing doesn't confirm the "$2T valuation" or "$750B compute" figures from press.
Anthropic has filed an S-1 registration statement with the SEC related to its proposed IPO.
Read more: anthropic.com/news/s1-sec
Legal disclaimer: A registration statement relating to these securities has been filed with the SEC but has not yet become effective. These securities may not be sold, nor may offers to buy be accepted, prior to the time the registration statement becomes effective. This post does not constitute an offer to sell or the solicitation of an offer to buy these securities, nor shall there be any sale of these securities in any state or jurisdiction in which such offer, solicitation or sale would be unlawful prior to registration or qualification under the securities laws of any such state or jurisdiction.
Joe Devon retweeted
Final version is out:
onlinelibrary.wiley.com/doi/…
"Reservoir Computing for Pattern Recognition and Time-Series Prediction Based on Nanofluidic Negative Differential Resistance"
Joe Devon retweeted
new #preprint! @niroshajmurugan
Does the brain know what's going on in a limb amputation wound - whether there's a pro-regenerative intervention or not? Let's look at its gene expression levels, with and without our biodome:
"Limb-localized induction of regeneration reshapes brain and amputation-site transcriptional states in adult Xenopus laevis"
biorxiv.org/content/10.64898…
yes, it knows. Also, lots of interesting transcriptional responses both in the limb and in the brain. Plenty of leads for subsequent investigation.
I was outvoted and we saw whalefall tonight.
By the end I was thinking I’d take 90 day fiancee over this, when will it end?
My whole family loved the movie. I was absolutely shocked.
There must be something wrong with me. I don’t get it.
It was painful beginning to end. One of the worst movies I’ve ever seen.
What am I missing?
Joe Devon retweeted
Common Crawl Foundation will be at IETF 127 in SF, week of November 15. If you are around, we'd love to chat!
Joe Devon retweeted
We've got an amazing group of people on here who care deeply about our products and want to hear from you. If you ever have feedback, tag them:
Claude Code
@addyosmani
@lydiahallie
@_catwu
@bcherny
@sidbid
@jarredsumner
@adocomplete
@delba_oliveira
@connerphillippi
@dickson_tsai
@dmwlff
Chat/Cowork
@felixrieseberg
@TheAmolAvasare
@RobertJBye
@sean_t_strong
Duet
@ahuja_karan
@gabemulley
Tag
@noahzweben
@ghodss
API/Platform
@RLanceMartin
@cjav_dev
@katelyn_lesse
@_hi_mc
@brada
General/Community
@edwinarbus
@sammcallister
@_sholtodouglas
(+ myself)
WHAT? Are you kidding me? This should result in jail time. How infuriating.
This slippery slope started with Google. They were a nice googley company. Fuzzy. Warm. Different than the others. We could trust them with ads in our gmail. They could be trusted to read our email.
We were concerned. But begrudgingly let it happen.
And now? Now all hell breaks out.
Joe Devon retweeted
Bingo.
I've worked in the water treatment & collection industry as well as the wastewater collection & treatment industry.
I can tell you right now no municipal physical public infrastructure is off the table here.
If you think I'm nuts, read about Stuxnet.
I've worked these treatment plants inside their SCADA systems & interconnected networks.
You've been warned, prepare accordingly.
Joe Devon retweeted
I released Surf CLI at the start of the year (9 days before Vercel's agent-browser) so agents can control my real browser. I still use it every day.
New update: Surf can fetch exact values off a page from a semantic description (like a flight departure time) without your agent reading the whole page. Cloudflare's Clairic finds each value by meaning rather than exact words.
github.com/nicobailon/surf-c…
Joe Devon retweeted
Yesterday was my last day at Anthropic.
I created Claude Code on desktop because I believe everyone can be a builder. Every feature request and bug report we got from this community was a gift that impacted millions of lives. I’m eternally grateful for those gifts, the product wouldn’t be what it is without you.
Anthropic is doing some of the most important work of our time. Contributing to that mission will forever be a career highlight for me. I never imagined I’d get to contribute to something so special in my lifetime.
Walking away from my best friends, a product I poured my heart into, and the mission was the hardest thing I’ve done in my career. But now I have some free time for GTA VI.
I'm already itching to get back to addressing all your feedback. I'll share what's next soon!