AI & Robotics News and Discussions

[firestar464]

Former OpenAI technology chief Mira Murati to raise capital for new AI startup, sources say

https://archive.ph/5r02w

[firestar464]

[firestar464]

DeepMind researchers find LLMs can serve as effective mediators

https://techxplore.com/news/2024-10-dee ... ctive.html

[Time_Traveller]

Ministers in AI ‘arms race’ with criminals and rogue states, senior MP warns

15 hours ago

Ministers and AI developers are locked in an “arms race” with criminals and rogue states attempting to use the technology for harm, a senior MP has warned, calling for more clarity on how it will be regulated.

Chi Onwurah, who chairs the science, innovation and technology committee, told The Independent that the “creativity and focus of criminal enterprises on adopting technology is unimaginable”.

But the Labour MP stressed that AI should not simply be seen as a force for harm, saying when used correctly it can be a “transformative technology that can really improve people’s lives.

“We really need to understand what it’s capable of, what it’s doing right now and the impact it’s going to have – and I don’t think we’re in that position.

“We’re clearly in an arms race with the criminals using AI, and those using AI for the public good… We need to make sure that the developers are on the side of the public good, because we know the criminals are absolutely focused on using it for criminal ends”, she said.

https://www.independent.co.uk/news/uk/p ... 30362.html

[firestar464]

Microsoft - "New autonomous agents scale your team like never before"

https://blogs.microsoft.com/blog/2024/1 ... er-before/

[firestar464]

User-friendly system makes it easier to verify an AI model's responses

https://techxplore.com/news/2024-10-use ... onses.html

[Time_Traveller]

US To Curb AI Investment In China Soon

October 22, 2024 12:50 pm IST

U.S. rules that will ban certain U.S. investments in artificial intelligence in China are under final review, according to a government posting, suggesting the restrictions are coming soon.

The rules, which will also require U.S. investors to notify the Treasury Department about some investments in AI and other sensitive technologies, stem from an executive order signed by President Joe Biden in August 2023 that aims to keep American investors' know-how from aiding China's military.

The final rules, which target outbound investment to China in AI, semiconductors and microelectronics and quantum computing, are under review at the Office of Management and Budget, the posting showed, which in the past has meant they will likely be released within the next week or so.

"It looks to me like they're trying to publish this before the election," said former Treasury official Laura Black, a lawyer at Akin Gump in Washington, referring to the Nov. 5 U.S. presidential election. Black added that the Treasury office overseeing the regulations generally provides at least a 30-day window before such regulations go into effect.

The Treasury Department published proposed rules with a raft of exceptions in June and gave the public a chance to comment. The draft rules placed the responsibility on U.S. individuals and companies to determine which transactions will be restricted.

https://www.ndtv.com/india-ai/us-to-cur ... on-6845498

[firestar464]

[firestar464]

Why I’m Leaving OpenAI and What I’m Doing Next

https://milesbrundage.substack.com/p/wh ... nd-what-im

Very insightful article.

I recognize that some will try to pigeon-hole me and my work in various ways no matter what – for example, given historical ties between the effective altruism, AI safety, and AI policy communities, I have lots of connections to the effective altruism community. And although those who know me know that I have also been a vocal critic of the community and don’t identify as “an EA,” I’m sure some will use these connections to dismiss me. Similarly, some may think I am too sympathetic to industry given my time there, or that I am otherwise biased in various ways – or they may draw similar conclusions based on who my future cofounders are. But I’d like to at least make such pigeon-holing be obviously unfair to anyone paying close attention, and I think being outside of industry will help towards that end.

Totally agree, weird techbro fanboys are annoying

[wjfox]

[firestar464]

Google unveils invisible ‘watermark’ for AI-generated text

https://www.nature.com/articles/d41586-024-03462-7

[Tadasuke]

Generative AI is driving many of the financial decisions in the tech sector right now, with companies like Google and Microsoft spending billions of dollars on powerful data center GPUs. As the generative models get larger and (allegedly) more capable, you need ever more GPUs to run them. Those chips don't last forever, though. A Google engineer has reportedly noted that data center GPUs have an expected lifespan of just three years, which could be very good for Nvidia and very bad for everyone else.

https://www.extremetech.com/computing/d ... -lifespans

[Yuli Ban]

Can't stress enough how cool this is to see, especially after a solid decade now since my big futurologist breakdown back in 2014 over ASIMO— and noticing quickly how every humanoid robot walked like it shat its nappy. I would always imagine the day we had androids that had an even remotely less awkward gait.
It's not actually that important, but it is cool!

[firestar464]

Google Preps AI That Takes Over Computers

https://www.theinformation.com/articles ... -computers

[weatheriscool]

Novel front-loading exoskeleton gets paralyzed patients up and walking
By Paul Ridden
October 28, 2024

Exoskeletons have been helping paraplegic patients to walk for years, but the latest development from KAIST is a little different. The WalkON Suit F1 can amble over to a patient seated in a wheelchair, wrap itself around the legs and body, and get the user up and walking.

Professor Kyoung-Chul Kong from KAIST's Department of Mechanical Engineering has been working on exoskeletons for around a decade, with the first WalkON Suit being revealed in 2016. By 2020, the research team had increased the walking speed to 3.2 km/h (1.98 mph) and entered the fourth-generation prototype into the Cybathlon in Switzerland.

This "cyborg Olympics" was launched in 2016 by ETH Zurich to drive advances in assistive technologies. The first event was held near Zurich and attracted teams from 25 nations, and saw 66 "pilots" compete for the spotlight. Four years later, pilots wearing KAIST's WalkON Suit 4.0 placed first and third in the Exoskeleton Race event. The third Cybathlon took place this past weekend, with Professor Kong's team fielding the latest development from the WalkON engineers.

https://newatlas.com/robotics/kaist-wal ... oskeleton/

[firestar464]

Based on everything we know so far from Jimmy:

(We were told that GPT-4.x (likely 4.5) is going to drop this month, though I'm very doubtful that we're suddenly going to be gifted a new model within the next 3 days. IDK anymore)

They probably gonna give ChatGPT a birthday gift sometime during or after Nov (that is, Dec), which is stated to be GPT-5 (aka Orion).
Anthropic's response to o1 (Claude 3.5 Opus?): December
(Likely something from OAI in response because as Jimmy said, it's doubtful Sam's gonna take his foot off the gas.)
Grok 3: December-January (o1 level)

[firestar464]

Whatever OAI drops in response to Grok 3 could very well be AGI based on current trends.

[firestar464]

New machine learning model can identify fake news sources more reliably

https://techxplore.com/news/2024-10-mac ... iably.html

[raklian]

[wjfox]

Can't stress enough how cool this is to see, especially after a solid decade now since my big futurologist breakdown back in 2014 over ASIMO— and noticing quickly how every humanoid robot walked like it shat its nappy. I would always imagine the day we had androids that had an even remotely less awkward gait.
It's not actually that important, but it is cool!

The clothed version is eerily human-like.