Emphyrio | Heurist | Zero1Labs | .skai

Introducing probably the most promising decentralised AI project on web3 : Zero1Labs. In this thread I will go briefly through all the information currently available on it z1labs.ai/early-access/?… #z1labs #zero1labs #DeAI #DecentralizedAI @zero1_labs

@cb_doge We don't need "legacy media" to know he is. I don't read legacy media but I've known he was a far right fascist and pro-neo-nazis just from reading his X posts back in fall 2024.

ELON MUSK: "What I've learned is that legacy media propaganda is very effective at making you believe things that aren't true."

@Matt_Pinner Cult-like cognitive dissonance. MAGAs are not *wrong*, they're just indoctrinated and can't accept truths. They need to be helped... You re not wrong guys, it's alright, you were never wrong, you were right to feel pain, you just listened to liars. Wake up.. unplug fox news..

what do you call it...

@PressSec Yeah, sure, we're going to trust your uneducated and uninformed propagandist lies echoing president Trump's over the opinion of **the head of national counter-terrorism**, a Trump loyalist with an actual conscience... 🤦

There are many false claims in this letter but let me address one specifically: that "Iran posed no imminent threat to our nation."   This is the same false claim that Democrats and some in the liberal media have been repeating over and over.   As President Trump has clearly and explicitly stated, he had strong and compelling evidence that Iran was going to attack the United States first.   This evidence was compiled from many sources and factors. President Trump would never make the decision to deploy military assets against a foreign adversary in a vacuum.   Iran is the world’s leading state sponsor of terrorism. The Iranian regime is evil. It proudly killed Americans, waged war against our country, and openly threatened us all the way up to the launch of Operation Epic Fury.   Iran was aggressively expanding their short-range ballistic missiles to combine with their naval assets to give themselves immunity – meaning they would have a degree of a capabilities that would give them immunity to hold us and the rest of the world hostage.   The regime aimed to use those ballistic missiles as a shield to continue achieving their ultimate goal – nuclear weapons.   The President, through his top negotiators, gave the regime every single possible opportunity to abandon this unacceptable course by permanently giving up their nuclear ambitions in exchange for sanctions relief, free nuclear fuel, and potential economic partnerships with our country.   But they would not say yes to peace because obtaining nuclear weapons was their fundamental goal.   President Trump ultimately made the determination that a joint attack with Israel would greatly reduce the risk to American lives that would come from a first strike by the terrorist Iranian regime and address this imminent threat to America’s national security interests.   All of this led to President Trump arriving at the determination that this military operation was necessary for U.S. national security, which is why he launched the massively successful Operation Epic Fury. The Commander-in-Chief determines what does and does not constitute a threat, because he is the one constitutionally empowered to do so - and because the American people went to the ballot box and entrusted him and him alone to make such final judgments. And finally, the absurd allegation that President Trump made this decision based on the influence of others, even foreign countries, is both insulting and laughable. President Trump has been remarkably consistent and has said for DECADES that Iran can NEVER possess a nuclear weapon. As someone who actually witnesses President Trump’s decision-making process on a daily basis, I can attest to the fact that he is always looking to do what’s in the best interest of the United States of America — period. America First.

@elonmusk Rofl that's the pot calling the kettle black! 😄😄😄🤦. Grok is atrociously unsafe... Not a single LLM is as poorly trained as Grok (even Venice which despite having no guardrails at all is way too dumb to be as dangerous as Grok)...

Keep ChatGPT away from kids and the mentally unwell

@bublboie @VraserX Ahah no need, a quick visit on your profile made the answer clear, my apologies for the doubts! ☺️ They were not fully unjustified though, as I am pretty sure that some people are honestly blind to the horrendous quality of such a video.

@EmphyrioTragedy @VraserX ask grok if i was being sarcastic

I just made a $500,000,000 AI movie in 24 hours. 100% AI. Hollywood is cooked. 🎬🤖

@bublboie @VraserX Hope you're sarcastic 😅. if this was the first minutes of a movie, I'd have already left the cinema and asked foe a refund.

@VraserX it looks so good and real everyone will watch this in theaters and love it this is the future

@aylacroft @krassenstein @sama Just to clarify, my post is refering precisely to the use of AI in national mass surveillance and in autonomous action *beyond* the PRU limits, which is what this Anthropic drama is about. Your linked posts don't contradict that. US is escalating further than anyone else atm.

@krassenstein @sama The answer is yes. I work in AI security & my work cited in their system cards. I also was in the military & married to a 1SGT. It's tagged as "national security " & unfortunately that will include us. It is really frustrating but we cannot be the only country not doing it atm

Tonight, we reached an agreement with the Department of War to deploy our models in their classified network. In all of our interactions, the DoW displayed a deep respect for safety and a desire to partner to achieve the best possible outcome. AI safety and wide distribution of benefits are the core of our mission. Two of our most important safety principles are prohibitions on domestic mass surveillance and human responsibility for the use of force, including for autonomous weapon systems. The DoW agrees with these principles, reflects them in law and policy, and we put them into our agreement. We also will build technical safeguards to ensure our models behave as they should, which the DoW also wanted. We will deploy FDEs to help with our models and to ensure their safety, we will deploy on cloud networks only. We are asking the DoW to offer these same terms to all AI companies, which in our opinion we think everyone should be willing to accept. We have expressed our strong desire to see things de-escalate away from legal and governmental actions and towards reasonable agreements. We remain committed to serve all of humanity as best we can. The world is a complicated, messy, and sometimes dangerous place.

@aylacroft @krassenstein @sama You're wrong about other countries doing it. Only Israel does. Russia doesn't have rhe means to, China is more ethical than what you would expect and actually passes new laws to prevent this. Europe.. is under US pressure&influence.. still has very strong laws to protect but..

@krassenstein @sama He just answered no, but that's PR lies. That would be incoherent : the DoD blacklisting Anthropic for maintaining these sageguards, then OpenAI signing with them hours later and pretending to maintain the same safeguards? Make it make sense, Altman.

@sama So will you allow your models to be used by the government for surveillance or autonomous weapons?

@sama Who do you think you're fooling, Altman? Anthropic is being blacklisted by the DoD (an EO has no power to change their name) for refusing to remove safeguards. You sign with the DoD *just hours later* and you pretend to maintain these same safeguards?? Lies.

@tszzl What's your take on Altman's choice to call it the DoW? It's still officially the DoD, EOs have no renaming power. Authoritarians try to rule by decrees, and playing along is.. capitulation to authoritarianism in my books.

most important part of this >We are asking the DoW to offer these same terms to all AI companies [...] We have expressed our strong desire to see things de-escalate away from legal and governmental actions and towards reasonable agreements.

@mark_k Sorry? Beung ethical is bad, in your books? Are you a psychopath, perhaps? Make it make sense 😵‍💫. I hope Dario will resist the pressure.

I've been warning people about Anthropic for years. Only now is it becoming obvious to everyone that not everything is good with that company. Quite the opposite.

@MarsUniversityX Psychopath defending empathy 😅👌

Elon Musk: "We want to have empathy that is deep, not shallow. I strongly believe we should care about humanity and we should care about the future. But we need to have empathy that is deep, not shallow Shallow empathy is caring about criminals, but deep empathy is caring about the victims of the criminals.”

@sama Oh and GPT5.2 got worse today. I prompted a short news (FAA, El Paso), it fact checked, confirmed & informed it was cancelled 3hrs l8ter, I said "good" & pondered possible links to Mexico → it went ballistic telling the FAA restriction didn't happen with adversarial tone...

We updated GPT-5.2 (the instant model) in ChatGPT today. Not a huge change, but hopefully you find it a little better.

@sama Well I unsubscribed, and now if I do log in rayher than using other brands' models, I quickly deplete my 10 prompts with 5.2 with nonsense demands so that I can use the ultra-cheap free model instead (much more pleasant...).

@Raw_0xDEADBEEF Amazing works, thanks!

Everyone says GPT-4o was better than GPT-5. We stopped guessing and measured it. 2,310 controlled comparisons. 8 judges. 5 providers. ~12,400 evaluations. Result: creativity dropped 6.7x. False refusal rose 4.4x. Benchmarks caught none of it. Paper: doi.org/10.5281/zenodo…

@KuittinenPetri @Raw_0xDEADBEEF Gpt-oss is easy to jailbreak though, compared to 5.2 which autocorrects like crazy. But even jailbroken, the creativity stays in the gutters.. it can only get rid of the false refusals (and the justified ones ^^).

@Raw_0xDEADBEEF With 900 million users OpenAI probably feels it needs to play safe, thus the excessive guardrails. The problem is that creativity & fun is sucked out of their newer models. Gpt-5.x is clearly inferior to gpt-4o in this respect and gpt-oss-120b is the worst offender. 1/

@tszzl Anyway I don't think there are chances to save OpenAI anymore. They had 4months to realize they were making some very serious mistakes.. they didn't. Now there's a horde of very active miscontent and OpenAI has apparently no clue of the terrible impact of negative advertisement.

Jesus on the mount: “Blessed are the poor in spirit, for theirs is the kingdom of heaven. Blessed are those who mourn, for they will be comforted—“ some guy in the audience: KEEP 4o

@tszzl Are you mentally deficient? Serious inquiry.. you certainly seem to be : you work for an enormous multi-billions company which relies in great part on public support and public image. Yet you mock part of that public, customers who express concerns, instead of trying to appease?

@sama Not much more to add. You already know why. And I *wasn't* in the 0.1%, because you had already made everything to dissuade from using 4o. You're going to lose much more than 0.1%

The 5.3 lovefest is so nice to see. Don't think we've had so much excitement for a model since the original GPT-4.