james snell

[james snell @provlynprotect]

A lab notebook is only corroborating evidence in a priority dispute — yet most researchers treat it as their strongest proof.

You wrote it; it sits on systems you control, so it can always be questioned. Support for your account is not proof on its own.

https://t.co/95sguAM4GV

[james snell @provlynprotect]

@technollama Landmark within Germany — but the reasoning turns on §44b UrhG, a German statute. No EU court has adopted this memorisation line yet, and a Hamburg court read the same provision the other way eight months earlier. Significant, but not continental.

[james snell @provlynprotect]

A German court just applied US copyright law.

The Munich Regional Court evaluated Suno's US training conduct under the four-factor fair use doctrine — and concluded it did not apply. Here is what the ruling actually decided.

https://t.co/uIaRHLPdxw

[james snell @provlynprotect]

Your commit history will not survive a patent dispute.

A commit timestamp is set by the machine that made it, and can be rewritten by a rebase or forced push. Opposing counsel knows this. Corroborative evidence isn't the same as verifiable proof.

https://t.co/GEuED0kVfc

[james snell @provlynprotect]

A free RFC 3161 timestamp is admissible in court. It does not carry a legal presumption of accuracy.

An eIDAS Article 41 certificate does. But it still leaves three gaps that most people do not consider.

https://t.co/fLbAHAdcNA

[james snell @provlynprotect]

Your IP insurance policy assumes you can prove what you own.

For patents and trademarks, a public register does that. For trade secrets, software, and creative works — there is no register. The evidence is whatever you made at the time.

https://t.co/9Cl3vKjMA7

[james snell @provlynprotect]

Most contract disputes aren’t about which deal was struck.

They’re about which version was approved — and whether that can be proved independently.

Every standard record of a negotiation is produced by a party with an interest in the outcome.

https://t.co/ZLjT4rLDjG

[james snell @provlynprotect]

@technollama The court's test assumes creators document their creative decisions as they make them. They don't — that's not how creativity works. The evidentiary gap isn't negligence, it's just the nature of the creative process. The protection has to be invisible to work

[james snell @provlynprotect]

@thebookseller The central challenge in every one of these cases is the same — proving what existed before it was ingested. Copyright is automatic. Proof of prior existence is not.

[james snell @provlynprotect]

@SpirosMargaris Data ownership is the right frame. But ownership without proof of prior existence is unenforceable. The question these cases keep returning to is not who owns it — it's who can prove they had it first.

[james snell @provlynprotect]

@technollama Whatever you think of academic publishers, the underlying question is the same — can you prove what existed, when, and in what form before it was ingested? That's a provenance problem

[james snell @provlynprotect]

VC disputes often come down to one question: were both sides reading the same version of the document?

A cryptographic fingerprint plus an eIDAS timestamped access log answers both halves of that question — immutably.

https://t.co/QYA3oBsu0V

[james snell @provlynprotect]

Founders share pitch decks with dozens of investors. Version history and access logs exist. What they cannot produce is cryptographic proof — hashed, qualified, and permanently anchored — in a single certificate a court can rely on.

https://t.co/XcVUUzkYJ6

[james snell @provlynprotect]

In a songwriting co-write, the question of who brought what into the room can take years to matter. By then, memory is all most writers have. Cryptographic prior proof, established before the session begins, is what changes the position.

https://t.co/HZzMXjZq7k

[james snell @provlynprotect]

UK copyright is automatic. But in a dispute, the burden of proof is entirely yours.

A file's creation date, an email, a social media post — none are independent evidence. Here is what the law actually requires.

https://t.co/CyE1CslBN5

[james snell @provlynprotect]

A US judge permanently dismissed xAI's trade secret case against OpenAI last month — not because the secrets didn't exist, but because they couldn't prove when they did. That evidentiary gap sits underneath almost every trade secret claim.

https://t.co/1MmQaOjjPP

[james snell @provlynprotect]

Five major publishers are suing Meta, alleging Llama was trained on 267TB of pirated content with copyright information stripped to hide the source. The case turns on provenance — who can prove what they created and when.

https://t.co/0suzgXv2nW

[james snell @provlynprotect]

@Ipkat A state office holding your documents for five years is a worse version of cryptographic proof of ownership. One requires you to remember to renew. The other is permanent by design

[james snell @provlynprotect]

Every consultant sends proposals and deliverables as digital documents. None of that produces an independent record of what each one contained when it left your hands. https://t.co/WvnL689E3v

[james snell @provlynprotect]

Cap table disputes are rarely about the cap table. They are about which version of a SAFE, note, or term sheet governed at the moment it changed hands. No standard tool produces an independent record of that. https://t.co/G87nAmsWGD