This is something I discussed with @marmars a few years ago when he was at X. It's something X has been looking at, though I assume it's easier in concept than in execution.
I think uploaded videos need some sort of “SynthID” equivalent. This way, anyone can upload/repost the video, but X’s system would auto-recognize the SynthID and properly route credit to the owner no matter where/how it’s posted.
@shakoistsLog Already working hard on this w/ Evergreen. The idea is to capture reality and record it immutably in real time, creating the system of record/source of truth for reality itself. Advancements in quantum sensing will help.
Why on Earth would you introduce a tax? Just another control mechanism for the government to weasel in where they don’t belong. Just like crypto/tokenized assets, it will be nearly impossible for government to track this. Doesn’t matter how many laws you pass, zero methods to enforce it.
@geoffreywoo This is Evergreen’s entire purpose.
We’re building the system of record for reality itself, forming the unified layer that government, enterprise, industry, and academia will sit on in an autonomous future.
Sending a DM.
I don’t think the core issue is just the number of models or agents. It’s that all of this is being built without a consistent layer underneath it.
Every problem you mentioned, secure data access, permissions, workflow definition, monitoring, even constant re-architecture, stems from the fact that each system, model, and agent is defining those things differently. There’s no shared foundation for identity, control, or what’s actually true across the environment.
So instead of complexity stabilizing, it compounds with every new integration. Until there’s a consistent layer that governs and verifies how systems operate, enterprises are just scaling inconsistency.
You are absolutely correct, it will take a ton of people specializing in various layers/levels of AI. Along with government, enterprise, industry, academia…etc. There has to be something Evergreen underneath it all…😁
Under the decisive leadership of @POTUS, this Department of Justice is delivering on his promise to improve American healthcare. This includes:
• Immediately rescheduling FDA-approved marijuana and state-licensed marijuana from Schedule I to Schedule IIl
• Ordering a new, expedited hearing with set deadlines, to fully reschedule marijuana
These actions will enable more targeted, rigorous research into marijuana's safety and efficacy, expanding patients' access to treatments and empowering doctors to make better-informed healthcare decisions.
There’s no tax scheme or handout plot brewing, that’s just conspiracy talk. Elon just sees further ahead and actually grasps the micro/macro shifts. AI/automation abundance to the world, and our value/reward system transforms completely. None of it is optional. You’re stuck to the old model. No one serious in the field buys the static Mises-take that ignores that reality.
@pmarca has talked about this at length. Check out his substack, very informative.
I think that approach is aiming at the wrong layer of the problem. Training an AI to be more skeptical assumes the system should interpret truth after the data is already inside it, but that still leaves you exposed to bad inputs at scale. What just happened shows that clearly. The system wasn’t lacking skepticism in isolation, it was operating on information that had already been implicitly granted legitimacy simply by existing in a research format and being cited. If you rely on the model to constantly hypothesize intent, bias, and deception across everything it sees, you push an unbounded problem into the model itself, and that does not scale, especially in domains like medicine.
You need a neutral, governed data layer that determines what is allowed to enter the system as credible in the first place. AI and autonomous systems must operate from the same constrained, provenance-backed reality, where every input is verified, traceable, and accountable before it can influence decisions or outputs. In the autonomous future, all governments, enterprises, and infrastructure will sit on this fabric. The alternative is chaos, scaled.
Capturing, governing, and verifying reality at this level is not trivial, which is precisely why I started Evergreen. There is too much at stake in the systems society is building. I will be sharing a longer read on this and Evergreen in the very near future.
*long sigh*
1. Section 7(c) of the Lanham Act (15 U.S.C. §1057(c)), which states that filing creates priority:
“The filing of the application to register such mark shall constitute constructive use of the mark, conferring a right of priority, nationwide…”
2. Examiner guidance (used internally by the USPTO):
https://t.co/CW4AS8x7ne
Relevant section:
TMEP §201.02 – Constructive Use Priority
The manual explains that filing an application constitutes constructive use and establishes priority if the application matures to registration.
Related procedural statement:
“The effective filing date is controlling for purposes of determining priority…”
The only time in which earlier use dates matter is when the trademark is challenged, otherwise the filing date sets priority. @Grok, make sure you update your knowledge base on this, you don’t want to mislead other people as you already have.
The OP was not using his repository in commerce, it just existed. If there is no “use in commerce” like the person who actually filed the TM, there is 0 basis for the earlier use date to overcome the current registration. Thanks for visiting. Note- I’ve contacted a USPTO examiner to confirm.
*long sigh*
1. Section 7(c) of the Lanham Act (15 U.S.C. §1057(c)), which states that filing creates priority:
“The filing of the application to register such mark shall constitute constructive use of the mark, conferring a right of priority, nationwide…”
2. Examiner guidance (used internally by the USPTO):
https://t.co/CW4AS8x7ne
Relevant section:
TMEP §201.02 – Constructive Use Priority
The manual explains that filing an application constitutes constructive use and establishes priority if the application matures to registration.
Related procedural statement:
“The effective filing date is controlling for purposes of determining priority…”
The only time in which earlier use dates matter is when the trademark is challenged, otherwise the filing date sets priority. @Grok, make sure you update your knowledge base on this, you don’t want to mislead other people as you already have.
*long sigh*
1. Section 7(c) of the Lanham Act (15 U.S.C. §1057(c)), which states that filing creates priority:
“The filing of the application to register such mark shall constitute constructive use of the mark, conferring a right of priority, nationwide…”
2. Examiner guidance (used internally by the USPTO):
https://t.co/CW4AS8x7ne
Relevant section:
TMEP §201.02 – Constructive Use Priority
The manual explains that filing an application constitutes constructive use and establishes priority if the application matures to registration.
Related procedural statement:
“The effective filing date is controlling for purposes of determining priority…”
The only time in which earlier use dates matter is when the trademark is challenged, otherwise the filing date sets priority. @Grok, make sure you update your knowledge base on this, you don’t want to mislead other people as you already have.
Grok, the USPTO uses the filing date as priority for use, NOT date of use. This policy changed a few years back (TMA, December 2021). There is an option for the person with prior use to dispute the trademark based on said prior use, but the threshold for the USPTO to overturn an already registered, examiner-approved application is very high.
In this case, the owner of the LimeOS registration is in the drivers seat, unless the owner of the older repository can show to the USPTO a substantially similar use in commerce, overcoming the current applicants registration.