ELON MUSK: AI competitors should peer review each other’s models before release. “I would recommend that we at least have some sort of informal weekly or biweekly call and that there’s maybe a week or two weeks of early access by competitors. This is why I think the incentives... work, competing companies are likely to understand if something is a security risk and they’re not going to be shy about highlighting that their competitors’ models should be delayed in release. The incentive structure does work quite well there. It’s similar to how the Motion Picture Association works as an industry group that decides how movies should be rated. That seems to have worked quite well. You really just have to ask the government to intervene if a company is doing something very dangerous and refuses to take action to reduce the danger.”show more

Knowledge already exists. What Elon is asking is: Who has the right to it? Because they're supposedly more responsible, or because they can pay for it?

Hahaha... Elon is worried that his Grok is lagging a bit behind Astra)

The motion picture ratings system doesn't work well. Look at what a Rated R or PG movie was back in 1968 with what one is now. As society changes, so does any grading system man sets up. What was considered an A when I was going to school is much easier to obtain now in many places in the US. Look at social media as an example with the "safety" committees. It's all arbitrary based on the plans of the few who are in control.

Basically, "your rivals will snitch on you for safety" and he thinks that's great. Honestly, might work.

Sehr gut, auch der Ansatz mit Selbstkontrolle - ich hoffe dass all die Klugen Menschen rund um KI dies zum Ansporn nehmen es wirklich als Chance zu begreifen🫡😉🥰

competition can move faster when accountability moves with it.

The joy is taking up more space than my to-do list. I'm happy to make a little fuss over this.

🤖🔥 This is an interesting idea. Competitors can often spot security risks that an internal team might miss. Independent peer review before major AI releases could add another layer of accountability. 🔍⚡

yeah pre release peer review sounds nice until competitors just bury each other tbh

Great idea! An AI rating system for competitors to review

Do you think China 🇨🇳 and Russia 🇷🇺 would pass that test?

You will never get Russia or China to do that...

Gbt loaded my raw data @elonmusk we both know that is not a good choice

Sampaikan saya tidak ingin melakukan INFORMAL AFAU FORMAL yang dimaksud, saya huga sudah Tidak mau kesana ... dengan maksud tujuan pakaii.. KECuALI AMBIL TAS.. atai keperluan lain. SAYA tidaj maI JIKA Ahirmya Seperti ini...

🔥

...not the worst idea. Models should peer-review other models as well.

The MPAA comparison is underrated. Movies didn’t wait for Congress to invent “PG-13.” The people who actually make the thing built the filter first. Same logic here: labs that can break a model are the ones who should see it before the public does

Boa ideia, mas duvido que as big techs topem abrir o jogo antes do lançamento.

🤔

Yeah we don’t need em being like the cats from an American tail and we are the mice

Can you make me an attachment for my current prescription lense glasses, that drops all hardware and tech excwpt the display and connection to my phone? I figured without the camera and speakers and probably something else in there it might look cool right now. I want an edc HUD

It’s high time Elon shall lay down the law on AI guidelines. Let’s not kid ourselves he’s the only one with ample knowledge and outlook.

Out handsome Genius Elon ❤️ is immensely right, 👏🏻 Great rest 💤😽

considering the greater risks, why not something more formal, rigorous with protocols and oversight like NIH clinical trials for experimental medications, etc ...publishing the open source results?...are red team guidelines applicable?

🤔

Weekly calls break once shipping cadence beats review speed. Pruning happens in the gap, not after.

🫶

"informal biweekly call" labs: sure let's schedule that for q4 2027