Harmony Hacker

The world is crazy and I'm just trying to make sense of it | Futurist | Creative Thinker | AI Enthusiast | Dog lover | Pro 🌍 | Pro Harmony | Anti Chaos

[Harmony Hacker @HarmonyHacker]

In retrospect I wonder if it would have made sense for Anthropic to never admit Mythos existed, and keep it as an internal edge until it was ready.

Seems the prospect of that IPO lead from one blunder to the next.

Also Fable 5 should have just been called Opus 5.

[Harmony Hacker @HarmonyHacker]

@davidpattersonx In another timeline they stayed in their lane, didn't play a highstakes game to slow down the rest of the AI field via doomerism... and just released Opus 5 with Fable 5 capabilities.

Right now they'd be working on Opus 5.1.

[Harmony Hacker @HarmonyHacker]

@ZackKorman @the_tech_space And yet one of them betrayed them to the government.

[Harmony Hacker @HarmonyHacker]

@davidpattersonx The only mistake Anthropic made is not calling Fable 5 just Opus 5.

The government is made up of cronies, crooks, and reactionaries (sometimes rolled up in one) and they will do as they will do.

Expecting 'good' from bad actors is likely what cost Anthropic here most of all.

[Harmony Hacker @HarmonyHacker]

@Flower_of_Babel @ChrissGPT @DavidSacks - Posted from China or Russia

[Harmony Hacker @HarmonyHacker]

They need to be fired? But there's no blame on the part of the administration speed-running banning a model in a single afternoon with barely any discussion with the team developing the model?

Perhaps the most insane factor here is that people are taking it at face granted that somehow the US administration has better AI researchers and better red-teamers than Anthropic. Wild.

And that they discovered a vulnerability in just 1 or 2 days.

The premise is beyond absurd. And people hop over points A, B, C, D and end up at Z blaming it entirely on Anthropic...

(Who btw, released the full model to numerous partners, including the government, specifically for testing and refinement)

Crazy stuff.

[Harmony Hacker @HarmonyHacker]

@liron Every word being spoken here is because they want Elon and SpaceX in Anthropic's place.

Since when did the all-in podcast become doomers?

[Harmony Hacker @HarmonyHacker]

Is this some sort of a joke? 'Now that the work has been done?'

It's the bleeding edge of tech. What work being done are you talking about? The model was released for just 3 days. We've barely just begun.

There's a handful of labs developing these types of models. And you're suggesting removing the developers on a whim... with an afternoon's worth of notice?

[Harmony Hacker @HarmonyHacker]

No, they actually can't 'fuck off' because that means fuck off to future advancements.

It also means fuck off to all foreign nationals working at all AI labs, as now there's a precedent that they can be considered persona non grata, and can be subsequently barred (on a whim) from the powerful models they're developing.

It's a disastrous message to send out to world star talent currently working in or being attracted to work at US labs. Wildly short-sighted. But that would be par for the course.

The administration is playing with fire with the one industry currently propping up the US economy.

[Harmony Hacker @HarmonyHacker]

@darasRantings @kimmonismus By wider impact you mean catastrophic impact.

The administration is playing with fire with the one industry propping up the US economy.

[Harmony Hacker @HarmonyHacker]

@dari1063 @ChrissGPT @DavidSacks The government banned employees who created the model from using the model internally.

That goes beyond export control.

It's a backdoor to an internal ban. As well as haphazardly threatening an entire industry that relies on foreign nationals as part of the backbone.

[Harmony Hacker @HarmonyHacker]

@Flower_of_Babel @ChrissGPT @DavidSacks The foreign nationals are the ones who **created** the model.

They are also the only ones currently capable of creating the model.

You're questioning their allegiances when you wouldn't be having this discussion without their vital contribution.

[Harmony Hacker @HarmonyHacker]

@scaling01 The problem isn't the admin saying whatever suits them...

...but that there is still a sizeable portion of the public that fails to recognize the most persistent and obvious patterns of deception.

[Harmony Hacker @HarmonyHacker]

@ChrissGPT The media can't help lying to the public

[Harmony Hacker @HarmonyHacker]

@DaveBlundin It's one of the most pivotal events because it's the US government stepping out and stating they can block the advancement towards AGI on a whim, or out of spite, or whatever nonsensical reason they hallucinate to fill in the gaps.

Low-key disastrous for the US tech sector.

[Harmony Hacker @HarmonyHacker]

@DaddyJimHQ @a_unbox @ai_for_success It's actually the least likely to hallucinate.

[Harmony Hacker @HarmonyHacker]

In case this isn't clear enough...

...the US government has neither competent nor more capable AI researchers compared to Anthropic.

[Harmony Hacker @HarmonyHacker]

@victormustar Correction: They would not be able to release GPT-6 because the administration drew a red line forbidding companies from developing models at Mythos-level or above.

[Harmony Hacker @HarmonyHacker]

@scaling01 And America will continue to be run by, for lack of a better term, short-sighted billionaire imbeciles.

[Harmony Hacker @HarmonyHacker]

She doesn't understand how the technology works. It's not a search engine 'scraping' and creating a database. It's not a database at all.

It's looking at the art and learning patterns from the art as a human would... except with far more efficiency.

It's not an 'art heist' if you go to the Louvre and learn to draw the Mona Lisa by staring at it for years and practicing daily.