[alancalverd]

There are always thousands of planes in the sky but never more than a handful moving on the ground at an airport. "All stop All stop All stop" prevents any collision on the ground, then you move them (and the ground traffic) one at a time.

Those in the sky do not require rapid servicing: they are sequenced at least two minutes, 3 miles, and 1000 feet apart if not in visual contact. Data  is passed upwards to initiate action, so "Contact Birmingham approach 131decimal 005 on handover" requires the pilot to enter and check the frequency of his next transmission: the readback (handshake) "Birmingham131005 for handover alfa charlie" confirms that a human flying XXXAC has done what was asked and understands that the approach controller is expecting him. No hurry.

A fine example of what can be done and said from a visual tower was given by a controller recounting a busy moment when a plane was accelerating along the runway. He said "Yankee 34 stop stop stop. Fire on port engine. Shut down and evacuate forward right side only. All traffic hold position. Fire truck is moving" It would take a pretty good AI system to notice the difference between normal "110%" exhaust flame and an uncontrolled blaze, never mind reassuring the captain that everything else is under control. What surprised the controller was that his pulse rate didn't change. 

[hamdani yusuf (OP)]

I am Confident About AI's Future - Open Source AI will WIN.

Open Source AI is the collective inheritance of humanity IMO

[hamdani yusuf (OP)]

"All stop All stop All stop" prevents any collision on the ground,

Unfortunately it doesn't work for airplane already flying.

[hamdani yusuf (OP)]

Worlds FIRST AGI SOFTWARE ENGINEER Just SHOCKED The ENTIRE INDUSTRY! (FULLY Autonomous AI AGENT)

Introducing Devin, the groundbreaking AI software engineer that's revolutionizing the field of coding and problem-solving. Devin is the new state-of-the-art on the SWE-Bench coding benchmark, showcasing its unparalleled ability to tackle real-world engineering challenges.

What sets Devin apart? This cutting-edge AI has successfully passed practical engineering interviews from top AI companies and has even completed real jobs on Upwork. Devin is a fully autonomous agent, equipped with its own shell, code editor, and web browser, enabling it to solve complex engineering tasks without human assistance.

But the true test of Devin's capabilities lies in the SWE-Bench benchmark, which evaluates an AI's ability to resolve GitHub issues found in real-world open-source projects. Devin's performance is nothing short of remarkable, correctly resolving an astonishing 13.86% of the issues unassisted. This far exceeds the previous state-of-the-art model performance of 1.96% unassisted and 4.80% assisted, setting a new standard in the field of AI software engineering.!

It's becoming increasingly important to know beforehand the end goal of AI agents before we give them the means to achieve it.

[hamdani yusuf (OP)]

Breakthrough in AI: Robots Now Learning on Their Own Like Humans - Ex-OpenAI

Covariant, a robotics company, is pioneering the use of AI similar to ChatGPT to create robots that learn and operate in the real world, revolutionizing industries like warehousing. Their advanced technology allows robots to understand and interact with their environment in ways previously unimaginable, handling tasks with human-like understanding. By blending digital data with sensory inputs, Covariant's robots represent a significant leap forward in making intelligent, adaptable machines that can perform complex tasks alongside humans.

[hamdani yusuf (OP)]

The First AI Virus Is Here!
The paper "ComPromptMized: Unleashing Zero-click Worms that Target GenAI-Powered Applications" is available here:
https://sites.google.com/view/compromptmized

TLDR

We created a computer worm that targets GenAI-powered applications and demonstrated it against GenAI-powered email assistants in two use cases (spamming and exfiltrating personal data), under two settings (black-box and white-box accesses), using two types of input data (text and images) and against three different GenAI models (Gemini Pro, ChatGPT 4.0, and LLaVA).

Abstract

In the past year, numerous companies have incorporated Generative AI (GenAI) capabilities into new and existing applications, forming interconnected Generative AI (GenAI) ecosystems consisting of semi/fully autonomous agents powered by GenAI services. While ongoing research highlighted risks associated with the GenAI layer of agents (e.g., dialog poisoning, privacy leakage, jailbreaking), a critical question emerges: Can attackers develop malware to exploit the GenAI component of an agent and launch cyber-attacks on the entire GenAI ecosystem?
This paper introduces Morris II, the first worm designed to target GenAI ecosystems through the use of adversarial self-replicating prompts. The study demonstrates that attackers can insert such prompts into inputs that, when processed by GenAI models, prompt the model to replicate the input as output (replication) and engage in malicious activities (payload). Additionally, these inputs compel the agent to deliver them (propagate) to new agents by exploiting the connectivity within the GenAI ecosystem. We demonstrate the application of Morris II against GenAI-powered email assistants in two use cases (spamming and exfiltrating personal data), under two settings (black-box and white-box accesses), using two types of input data (text and images). The worm is tested against three different GenAI models (Gemini Pro, ChatGPT 4.0, and LLaVA), and various factors (e.g., propagation rate, replication, malicious activity) influencing the performance of the worm are evaluated.

As the machines become more conscious, they will need their own version of moral standard they can follow and apply.

[hamdani yusuf (OP)]

Breakthrough in AI: Robots Now Learning on Their Own Like Humans - Ex-OpenAI

Covariant, a robotics company, is pioneering the use of AI similar to ChatGPT to create robots that learn and operate in the real world, revolutionizing industries like warehousing. Their advanced technology allows robots to understand and interact with their environment in ways previously unimaginable, handling tasks with human-like understanding. By blending digital data with sensory inputs, Covariant's robots represent a significant leap forward in making intelligent, adaptable machines that can perform complex tasks alongside humans.

[alancalverd]

Unfortunately it doesn't work for airplane already flying.

As I mentioned in the next paragraph, airborne collisions are prevented by separation.

[hamdani yusuf (OP)]

Unfortunately it doesn't work for airplane already flying.

As I mentioned in the next paragraph, airborne collisions are prevented by separation.

Afaik, humans can only perform well for up to 5 tasks at once. They are not especially good at multitasking.

[alancalverd]

They say that flying a helicopter requires two brains and three hands, but some fairly ordinary people do it for a living. And the object of air traffic control is to eliminate some of the tasks associated with operating under visual flight rules. 

[hamdani yusuf (OP)]

AI Agents Take the Wheel: Devin, SIMA, Figure 01 and The Future of Jobs

Devin, SIMA, Figure 01, all in 24 hours. What does it mean and are AI models taking the wheel? I?ll go through 5 relevant papers and 11 articles to get you all the relevant details, from what exactly Devin accomplished, and didn?t, to DeepMind's new AGI-attempt-in-3D (SIMA) to just how far AI agents have come and what that means for the future of jobs.

As the machines become more conscious, they will need their own version of moral standard they can follow and apply.

OpenAI Deploys AGI Into Humanoid Robot - Displays STUNNING Abilities (Figure 01 Breakthrough)

Figure 01 gave another incredible update on their progress. Their robot can now have entire conversations powered by ChatGPT. Plus, we look at other incredible robots making great progress.

[hamdani yusuf (OP)]

AGI is Already Here: SHOCKING Details Exposed

What is AGI? What does it take to achieve AGI? What are the levels of AGI?

21:58 Leaked Document

This is just a leak, but sounds believable.

[hamdani yusuf (OP)]

Nvidia Reveals Omniverse Cloud Streams to the Vision Pro

Watch Nvidia CEO Jensen Huang show how Nvidia's Omniverse Cloud streams to Apple's Vision Pro XR headset.

Nvidia Reveals Project GROOT and Disney Bots at GTC Conference

Nvidia CEO Jensen Huang shows new robot technology at its GTC conference in San Jose.

NVIDIA Robotics: A Journey From AVs to Humanoids

See NVIDIA?s journey from pioneering advanced autonomous vehicle hardware and simulation tools to accelerated perception and manipulation for autonomous mobile robots and industrial arms, culminating in the next wave of cutting-edge AI for humanoid robots.

Experience our journey from simulation to real-world deployment, showcasing our commitment to innovation and technological excellence.

[hamdani yusuf (OP)]

Nvidia 2024 AI Event: Everything Revealed in 16 Minutes

Nvidia CEO Jensen Huang kicks off its GTC keynote in San Jose with a slew of AI infused chip announcements. Check out our recap right here.

[hamdani yusuf (OP)]

Humans interaction with virtual universe will be significantly improved, at least in terms of bandwidth.

Elon Musk Reveals His STUNNING Human Neuralink Patient | The Brain Computer Interface N1

00:00 First Neuralink Patient Shares His Results
08:50 Commentary

[hamdani yusuf (OP)]

STUNNING Breakthrough "AGI Robot" From OpenAI, 1x, NVIDIA, Boston Dynamics, Anduril

Chapters:
0:00 - Intro
0:24 - 1X EVE
3:02 - Project GROOT
10:17 - Boston Dynamics
13:07 - Mercedes Robot
14:17 - Xiomai Dog Robot
14:50 - Yondu Robot
16:27 - Anduril Warfair Robot

Training in a virtual universe has a lot of benefits, such as time, energy, and money. But in the end the robots need to fine tuned by real world situations, although it's more likely that sooner or later they will get the capability to learn for themselves.

[hamdani yusuf (OP)]

Anon Leaks NEW Details About Q* | "This is AGI"

An interesting comment from the video.

It sounds like Q* is an upgrade from the greedy approach of LLMs where they only finding the highest probability of the next token in the answer, to finding the the highest  probability of all the tokens put together. With my limited understanding in this, it sounds like they're accomplishing this with having a second latent space. So we basically go from
a normal LLM: input-text -> latent space -> output-text,
to
Q*: input-text -> input latent space 1 -> latent space 2 (i.e. EBM) -> output latent space 1 -> output text.

We might finally get an LLM that can answer the age-old question of "How many tokens are there in your response"

[hamdani yusuf (OP)]

Ai and education. The end of schools as we know them. When AI can understand all humans and computer languages, what we need to do is specifying our goals and targets.

[alancalverd]

When AI can understand all humans

Probably an extreme case, but whenever I've worked  in the field of disability and rehabilitation we have been faced with the infinite variability of human ability, multiplied by the infinite variability of human response to any interaction. Whatever you make or do for one patient, there is always somebody for whom it's not quite right and/or who wants it in a different color. To a lesser extent, teaching is the same.

Now the problem you have demonstrated many times in these forums is that AI can only generate outputs based on previously published data, most of dubious value, and a learned or programmed preconception of acceptable phraseology. This puts it in the category of "educationalist" rather than "teacher."

The difference is

An educationalist takes a subject that he doesn't understand, rephrases it so that nobody can understand it, and insists that this must become part of a curriculum because all pupils are identical.

A teacher takes a subject that he understands, and rephrases it so that a child can understand it. The curriculum is less important than the pupil, and all the pupils are different.

[hamdani yusuf (OP)]

Probably an extreme case, but whenever I've worked  in the field of disability and rehabilitation we have been faced with the infinite variability of human ability, multiplied by the infinite variability of human response to any interaction. Whatever you make or do for one patient, there is always somebody for whom it's not quite right and/or who wants it in a different color. To a lesser extent, teaching is the same.

I was referring to languages.