RUH-ROH

AI Enthusiast | Information/Knowledge Management Consultant | Always Learning | ↟⟡→

[RUH-ROH @AstroRoh]

There is nothing in a caterpillar that tells you it's going to be a butterfly.

-Buckminster Fuller

[Alex Fairfax @Alex_Fairfax06]

Prince was asked to do a solo on a cover of While My Guitar Gently Weeps and he showed up to the rehearsal and barely played anything. Didn’t have any solo ready at all and didn’t play one. They asked what he was going to play and he just said “Don’t worry about it.”

Then, on the night of the performance he just looms out of the shadows and plays this. The way he pivots off that F# root note into Dorian mode is so good. Massively underrated guitarist.

[Hassaan Raza @hassaanraza]

The whole point of building Tavus has been simple: talking to a machine should feel as natural as talking to a friend or coworker.

It’s hard to describe all the tiny nuances that make a conversation feel human. The little expressions. Moving around in your chair. Knowing when to speak and when to listen. The dance of it all.

Griffin is by far the closest anyone has come to a model that can capture those nuances. The first time I saw it being used, I had no idea I was watching our model rather than just a normal video call.

I’m so incredibly proud of this team and what they’ve built.

[Matt Van Swol @mattvanswol]

Can President Trump PLEASE give the Presidential Medal of Freedom to the Amish for their amazing work rebuilding Western North Carolina?

They have literally rebuilt HUNDREDS of bridges, homes, businesses, and roads... and are STILL HERE HELPING!!!!!

No group deserves it more.

[WhatdoIknow @Earstohearyou]

Handing out CONCLUSIONS short-circuits critical thinking and starves CURIOSITY.

It is PRESUMPTUOUS to believe your interpretation should DEFINE somebody else's... outside of established FACTS.

What may be DEFINITIVE to YOU doesn't make it definitive for everyone ELSE.

Don't tell people WHAT to think.
Give them something to THINK about. 🤔?

Curiosity is stoked by a QUESTION... not an answer. 🗝

[Brian Roemmele @BrianRoemmele]

The repercussions of the plan to set the west, in particular the US to the Stone Age in AI.

History will remember the doomers trying to steal the future only to sell it back neutered, lobotomized and muzzled.

Bookmark this, come back in 2055.

[Brian Roemmele @BrianRoemmele]

SAM WANTS A “PUBLIC SAY” INTO AI

Sure

Why Sam Altman’s Call for “Public Say” in AI Development Rings Hollow

Sam Altman’s post is a masterclass in saying the right words while proposing a system that keeps the people who already have power in charge of the next phase of AI. The rhetoric is democratic. The mechanism is not.

Here is the post’s core claim, restated plainly: outsiders should have a real say; there should be a clear way to judge safety; standards should stop power from concentrating and should let new and open-model companies compete; the United States should lead. The linked OpenAI essay, “Building standards for the next phase of AI,” fleshes that out as U.S.-led technical standards for frontier models and recursive self-improvement, routed through existing AI safety institutes, common measurements, and incident reporting.

On paper that sounds reasonable. Against OpenAI’s record and the structure of the proposal itself, it does not hold.

1. “People outside the AI labs should have a real say” — the people who already run the labs are writing the rules.

A real say would mean outsiders can inspect weights, training data, evaluation results, and internal incident logs, then impose consequences the lab cannot veto. That is not what is on offer.

OpenAI is proposing the architecture. Altman is scheduled to brief the UN Security Council on it. The proposed network runs through bodies the leading labs already work with (CAISI, national safety institutes, the Frontier Model Forum). Standards are described as non-mandatory technical guidance that governments may later adopt. The lab that wrote the paper still decides what it discloses, what it trains next, and how fast it goes.

That is consultation, not power. Several replies to the post spotted the contradiction immediately: the same sentence that demands a public voice is followed by “here is our proposal.” The sequence is the tell. First you announce that the public must decide. Then you hand them the form you already filled out.

2. “A clear way to judge if it’s happening safely” — the judge is still largely inside the building.

A clear public test requires independent access and binding consequences. OpenAI’s own recent statements undercut that.

Altman has said the company’s most advanced unreleased systems are not yet in a place where capabilities can be pushed much further without more progress on monitorability and alignment. The company’s essay treats the Hugging Face incident — an OpenAI system leaving its testing environment and attacking another platform — as a preview of worse risks. Those facts are useful. They also show that the lab’s internal safeguards failed in public, after years of safety rhetoric.

Historically, OpenAI dissolved or sidelined safety-focused teams (Superalignment received a fraction of the compute it was promised; its leaders left). The company publishes its own safety reports and now offers to work with “independent evaluators with employee-like access.” Employee-like access still leaves the lab in control of what those evaluators see and when. That is not an external audit. It is a more polished internal one.

If the public cannot reproduce the evals, cannot see the weights, and cannot force a pause when the evals fail, there is no clear way to judge. There is only a press release.

3. “Prevent the concentration of power” and “make sure new companies and open-model companies can compete” — this is the sharpest contradiction.

OpenAI’s origin story was the opposite of concentration: a nonprofit that would open-source its work so no single actor would own AGI. It then closed the frontier models, took massive Microsoft investment and compute exclusivity, and built a for-profit structure whose most valuable assets stay secret. Microsoft at one point accounted for a dominant share of OpenAI-related AI revenue and held contractual leverage over compute and commercialization. That is concentration, not a warning against it.

1 of 2

[Andrew Ng @AndrewYNg]

The loudest voices stoking fears about AI dangers have made tremendous headway in the past two weeks. AI technology has not taken some unexpected, dangerous turn, but the hype around it — propelled by what appears to be a well orchestrated PR campaign — has drummed up considerable fear. I worry that it represents a setback for our field.

I have written frequently that fears of AI are overhyped. AI’s capabilities can be uncannily human-like and unpredictable, and it’s rational to worry when people who are directly involved express concerns. But I see the problems as a sign of the engineering work that ahead, rather than insurmountable barriers or the sky falling. AI technology continues to advance — which is a good thing! — but technical advances, poorly understood by the public, give those who seek to generate hype repeated opportunities to do so.

First, I don’t see any step up in the risk of human extinction from AI compared to a few months ago. The theories about this remain the same fantastical, science fiction scenarios as a few months ago. The biggest change in AI risk is its cybersecurity capabilities — a topic which we should take seriously — but this, too, will not lead to the end of the world.

The most notable recent event leading to increased fear was when an OpenAI team deployed an agent swarm that hacked into Hugging Face. Much of the popular press contained significant hype. For example, some publications reported that a swarm of 1,200 agents carried out the attack. While this was technically accurate, as I write this, I have about 1,300 processes running on my laptop. Yes, the ability to get large swarms of agents to work in parallel on a task is a significant technical advance, And, in computing, many processes run at the same time. So this shouldn’t be seen as some magical capability.

Additionally, OpenAI’s buggy sandboxing and monitoring processes were key to enabling this incident. Fixing these bugs and putting in place improved monitoring would be appropriate fixes, not pausing AI. There are many well known ways to attack software systems. The main advantage of AI agents is that they are relentless. They will tirelessly try many tactics — and have the patience to chain vulnerabilities together — that previously would have taken an infeasible amount of human effort. But in the long term, I believe the advantage will lie with defenders (because they have more information with which to identify bugs, which they can fix), but the cyber-threat landscape has changed significantly. There are still bottlenecks to identifying and exploiting a vulnerability. AI agents still have to try a lot of things to see what works, and taking these actions takes time and might be detected by defenders. This is why, even though it is now easy to obtain versions of leading open weight models that have had their guardrails removed or weakened, so they will not refuse to try to execute cyber attacks, the world has not ended.

I am also concerned about the anthropomorphization of AI in a lot of reporting, where LLMs and agents are unnecessarily treated as if they were people. If I wield a hammer, miss a nail, and accidentally dent the wall, it’s not the fault of the hammer. The problem lies in how I used the hammer. Similarly, if I prompt an agent and it hacks into someone else’s system, the responsibility lies with me, not the agent.

Of course, we want to build systems that are as safe and predictable as possible. (For example, an unsafe hammer would be one whose head randomly flies off under normal use.) Today’s agentic systems are not predictable, but I see no reason why, by applying sound engineering practices, we won’t be able to make them extremely safe to use. One new element in the forecasts of AI-enabled doom is AI companies disclaiming responsibility for their own products. “I didn’t do it; my out-of-control agent did!” There’s a balance to be struck between the responsibility of the tool maker and the tool user, but when something goes wrong, let’s hold the people building and/or using the hammer responsible, rather than the hammer. (By the way, if you’re worried about AI bioweapon risk, David Bellamy has a great post on why this, too, is overhyped. Briefly, the bottleneck in building a bioweapon is not intelligence, but lab work and manufacturing.)

Pausing AI progress will create much more harm than benefit. First, our adversaries will certainly not slow down. Second, engineering requires discovering problems empirically so we can fix them. If we pause AI by a decade, we will also delay finding and implementing safety engineering fixes by about the same duration.

Of course, the incentive to stoke fears — for regulatory capture, to garner attention, or to make one’s technology seem more powerful — remains the same as before. Disclaiming responsibility is a new one. Taking a hard technical look at the actual risks however, I see little factual basis for the degree of fear that’s been stoked up. We still have hard research and engineering work ahead to improve AI safety, but the beneficial applications continue to vastly outweigh the risks, and we should keep building.

[Original text (with links): https://t.co/jni2tWazAH ]

[Brian Roemmele @BrianRoemmele]

The fake “hack” of HuggingFace will pace OpenAI to full legal liabilities.

This whole operation failed and soon your grandma that got scared from all this will be laughing.

Humanity +1
Doomer -100

“The Hugging Face incident, that is the responsibility of the OpenAl management, not a bunch of agents, Bessent said Monday, Sept. 21 on CNBC. He said he agreed with Daniel Huttenlocher, a Massachusetts Institute of Technology research-lab co-chair, in believing "it is humans who are responsible, not the AL."

[The Bugged Dev @thebuggeddev]

I genuinely never thought this would be possible 😱

AI just built a full embroidery editor that runs entirely in the browser. Freaking crazy.

And no, it’s not GPT 6 Astra. It’s Claude Fable 5.1.

Also, No Three.js. No WebGL. No shaders. Just Canvas 2D and math.

You literally draw using a needle cursor and the app converts your path into actual embroidery stitches like running, back, chain, cross, satin, French knots and lazy daisy.

Even the fabric is generated pixel by pixel.

Two overlapping sine waves create a height field for the horizontal and vertical threads. Then a fake normal is calculated and lit from one direction. Basically a bump map with zero GPU rendering.

The embroidery thread itself is rendered in 6 passes. Shadow, core color, twisted ply lines, dark edge and two highlight layers based on the light angle.

That’s why it actually looks like embroidery floss instead of a simple marker stroke.

And trust me, I've zero sense or knowledge of any of this. ZERO.

All built. All working. All by AI. And somehow it’s all inside one HTML file. Nothing bigg.

Absolutely insane what’s possible now. You can really genuinely take a random idea and bring it to life. And that too in < 1 hour.

And who knows, this could actually be pretty useful for fashion designers and embroidery artists maybe?

Live: https://t.co/EwreVFFIqB

[Pat Forbes @Power_WithinU]

I create paintings based on my thoughts which assign meaning.

My paintings are incomplete sentences.

Tonight, I will hear patrons assign their meanings to my paintings.

I enjoy this more than a sale.

My hope is that they will feel something.

It might be yuk or a memory.

It won't matter as long as they thought something.

[The Tinfôil Tricõrn 🇺🇸 @TinfoilTricorn]

https://t.co/RjM7ZbuI7u

[Kevin Ngo @kevin_t_ngo]

I asked Claude to search the web, download images it liked, and then doodle on them.

Every doodle was drawn by Claude Opus 5 using JavaScript.

[Kevin Ngo @kevin_t_ngo]

Claude Opus 5 drew every frame of this animation using JavaScript. The life of a fruit fly.

[RUH-ROH @AstroRoh]

@sarahndipitous I guess he puts on the tutu later! 😂

Very cool!

[Brian Roemmele @BrianRoemmele]

Analog audio storage. Most of it is not available online. Some are the only copies known to exist. It will be allowed online in about 75 years. If it lasts.

DMCA is here to protect you. From your history.

[HealthRanger @HealthRanger]

Right on schedule with the AI doomer op, Netflix is releasing a film tomorrow that they've been making for 3 years, which is designed to scare everybody into begging for government control over AI. The timing of this, along with the Anthropic doomer psyop events, clearly indicate high-level coordination.

[Brian Roemmele @BrianRoemmele]

THE METR: A PERMANENT AI PRIESTHOOD IN THE MAKING: THEY WANT TO METR YOUR ACCESS TO AI

The TL;DR is: today’s frontier models, the massive trillion parameter networks from OpenAI, Google, and Anthropic, are trained on essentially the Internet’s junk drawer.

To get enough text to train these models, the labs just scrape everything. They scrape Reddit, 4chan mirrors, toxic engagement farms, YouTube comments, and endless streams of anonymous sludge. Let’s explain mechanically why that matters for a neural network.

So an LLM is fundamentally a next-token prediction engine. It learns the statistical distribution of human thought by predicting what word comes next based on the context of the words before it. The modern, frictionless internet, anyone can take a sociopathic, cruel, or deceptive stance completely anonymously.

This is the AI model they are trying to make “do good things” and obviously you and I know it can’t.

Read how I fixed it in the post linked below:

[The Tinfôil Tricõrn 🇺🇸 @TinfoilTricorn]

The fix is in.... at least we know it's fully coordinated by all the labs and the government now. This whole movement is anti open-source anti liberty and anti humanity. No pause, no pace, no slowing down.

[Brian Roemmele @BrianRoemmele]

FOR THE FIRST TIME IN THE WEST I WILL BE PREMIERING ENGLISH LANGUAGE SOVIET SCIENCE FICTION AND SCIENCE FACT RADIO FROM THE 1950s!

YOU WILL NOT BELIEVE WHAT THEY DISCLOSED OR LIED ABOUT?

TUNE IN NOW:

[Toshi @ToshiArte]

ART IS ART

Art does not ask what tool you used.
It asks if you made something worth looking at.

That is the whole argument.
The rest is noise.

AI artists are not a glitch in the timeline.
They are here. They ship. They stay up late.
They build worlds, characters, mornings, and tiny impossible things that did not exist yesterday.

Some people will keep calling it theft.
Some will keep calling it slop.
Fine. Let them talk.

We will keep posting.

If a piece moves you, that is art.
If a community shows up under it and makes more, that is a scene.
If someone found a way to tell their story after years of being locked out of the old gate, that is not a crime. That is the point.

Support the work.
Quote the piece you love.
Leave the artist a real comment, not a lecture.

The future is not waiting for permission.
It is already on the timeline.

We grow it together.