AI & Robotics News and Discussions

[funkervogt]

This excellent essay on Astral Codex Ten explains why the success of GPT-5 is so important to the pace of AI development. Let me explain:

There have been four GPT AI's so far: GPT-1, 2, 3, and 4. On average, each took 18 months to invent after the last.

Each one cost roughly 30x as much money to create as the last. Because GPT-4 cost $100 million, that means GPT-5 will cost about $2.5 billion.

The company that created the GPT machines, OpenAI, has enough money to make GPT-5, and it will probably be released in the next year. However, the cost trend means GPT-6 will cost $75 billion, which is well beyond its means.

Microsoft owns OpenAI, and if GPT-5 is a blockbuster product (e.g. - eerily smart and competent), Microsoft could, with difficulty, make the $75 billion investment to build GPT-6.

The cost trend also reveals that GPT-7 will cost $2 trillion to build, which no one has to spare. Only a major country or group of countries could afford it, and they'd only spend that much if GPT-6 were a super-blockbuster product, making the reward of an even better computer worth it.

The amount of computer chips and electricity needed to support each successive GPT machine has also been increasing exponentially, and while building GPT-5 is still within OpenAI's means, GPT-6 isn't--it would need something like a town-sized server farm and a dedicated nuclear reactor.

So if GPT-5 is underwhelming, isn't massively adopted across the world for all kinds of tasks, and doesn't make OpenAI billions of dollars in profit per year, it won't be worth it for anyone to build GPT-6 at the same pace the previous five GPTs were built. Instead, they'll slow down and build GPT-5.1, then -5.2, etc., with a year or two between each iteration. Moore's Law and variety of small hardware and software improvements will make a slower development strategy affordable: by the time GPT-6 capability is achieved, the total cost will have been much less than the $75 billion that would have been needed to build it a decade sooner and in one, big leap.

If the scenario in the preceding paragraph materializes, the general public will misperceive it as "proof" that AI was hype and can't actually be built, in the same way that the slowdown in progress with autonomous vehicles has convinced people the technology will never succeed.

The scaling problem with LLMs hits home how bad their learning algorithms are, and how much the tech industry is trying to brute force its way to creating an AI by throwing more computer chips and electricity at the problem. Within the next three or four years, we'll know whether this paradigm can continue, and there will be major implications for the tech industry and popular perceptions of "AI."

If GPT-5 is really good, then buckle up because mass job displacements, social changes, science and tech acceleration, and Manhattan Project-levels of investment in AI that will be felt across politics and the world economy will start before the end of this decade.

However, I predict it's much likelier GPT-5 will fall below the necessary performance threshold. While an impressive product, it won't be good enough to inspire the immediate creation of GPT-6. LLM's will probably suffer from an insurmountable "last mile problem" that keeps them from evolving into true AGIs, though in many narrow domains they will equal or surpass human intelligence and competency and they'll be able to convincingly "fake" general intelligence 99% of the time (which some dumber humans will mistake for proof GPT-5 is sentient). I believe we'll have to wait for several important algorithmic breakthroughs to happen to achieve true AGI, which I think could happen as early as 2050.

https://www.astralcodexten.com/p/sam-al ... 7-trillion

[Time_Traveller]

AI could 'supercharge' election disinformation, US tells the BBC

2 hours ago

Artificial intelligence (AI) threatens to "supercharge" disinformation and incite violence at elections, the US deputy attorney general has warned.

Speaking exclusively to the BBC, Lisa Monaco described AI as the "ultimate double-edged sword".

It could deliver "profound benefits" to society but also be used by "malicious actors" to "sow chaos", she added.

And she revealed plans to make the use of AI by criminals an aggravating factor in sentencing in US courts.

The former federal prosecutor, who is in the UK to deliver a lecture on AI at the University of Oxford, said violent criminals who used guns were given longer sentences.

https://www.bbc.co.uk/news/world-68295845

[weatheriscool]

Amazon unveils largest text-to-speech model ever made
https://techxplore.com/news/2024-02-ama ... peech.html
by Bob Yirka , Tech Xplore

A team of artificial intelligence researchers at Amazon AGI announced the development of what they are describing as the largest text-to-speech model ever made. By largest, they mean having the most parameters and using the largest training dataset. They have published a paper on the arXiv preprint server describing how the model was developed and trained.

LLMs like ChatGPT have gained attention for their human-like ability to answer questions intelligently and create high-level documents. But AI is still making its way into other mainstream applications, as well. In this new effort, the researchers attempted to improve the ability of a text-to-speech application by increasing its number of parameters and adding to its training base.

The new model, called Big Adaptive Streamable TTS with Emergent abilities, (BASE TTS for short) has 980 million parameters and was trained using 100,000 hours of recorded speech (found on public sites), most of which was in English. The team also gave it examples of spoken words and phrases in other languages to allow the model to correctly pronounce well-known phrases when it encounters them—"au contraire," for example, or "adios, amigo."

[weatheriscool]

The Seeing Eye Dog V2.0 is shaping up as a gamechanger
By Mike Hanlon
February 16, 2024

https://newatlas.com/robotics/seeing-ey ... mechanger/

An important new iteration of the robotic seeing eye dog was shown to the world this week, when Glasgow University showed off RoboGuide, an AI-powered quadruped for visually impaired people. The future for robotic canine assistants looks not just commercially huge, but also massively empowering for the world's one-third of billion visually impaired people.

Replacing a real "seeing eye dog" with a robot guide dog might seem like a step backwards, particularly considering the life-affirming nature of the inter-species collaboration and the obvious emotional bond.

That's until you realize that only one in every 15,000 visually impaired people has a service dog of any kind.

[Tadasuke]

This good article available in text and voice form is very relevant to this AI & Robotics topic: https://humanprogress.org/ai-tractors-a ... g-devices/

[spryfusion]

[firestar464]

robot do kid drawing

https://techxplore.com/news/2024-02-hum ... robot.html

[weatheriscool]

Nvidia, OpenAI, Microsoft, Other Tech Titans Investing In Humanoid Robot Startup
Figure AI aims to put its robot, Figure 01, in high-demand and physically dangerous roles.
By Adrianna Nine February 27, 2024
https://www.extremetech.com/computing/n ... ot-startup

What do Intel, LG, Nvidia, Microsoft, Samsung, and OpenAI currently have in common? They’re all taking big bets on Figure AI, a startup focused entirely on building and deploying humanoid robots. The California-based company has reportedly secured $675 million in funding, having attracted some of the most powerful entities within the tech industry.

According to Bloomberg, Microsoft has invested $95 million in Figure AI, while Nvidia has committed $50 million. Intel Corp.’s venture capital division has invested $25 million; LG Innotek $8.5 million, Samsung $5 million. Jeff Bezos has committed a whopping $100 million, and OpenAI, after noodling on acquiring the startup and deciding against it, has instead devoted $5 million to Figure AI’s cause. Venture funds from Parkway Venture Capital to Tamarack are also helping Figure AI surpass its original $500 million funding goal.

[firestar464]

New AI model could streamline operations in a robotic warehouse

https://techxplore.com/news/2024-02-ai- ... house.html

[weatheriscool]

Electric Sheep's jolly AI-powered garden bot edges, trims and blows
By Paul Ridden
February 29, 2024
https://newatlas.com/robotics/electric- ... nce-robot/

Robotic garden helpers have been keeping lawns in trim for decades, but none are as much fun to watch as the Verdie AI-powered outdoor maintenance bot from Electric Sheep as it wheels around edges, blows debris and gets to grips with power tools.

For the last few years, San Francisco's Electric Sheep Robotics has been providing automated outdoor maintenance tech as a Robots-as-a-Service rental model. But more recently the company has been acquiring traditional outdoor service providers and "progressively transforming operations by deploying proprietary AI software and robots" in a move towards becoming a large-scale outdoor maintenance company.

Tapping into this US$1 trillion market involves replacing the "wide variety of highly pollutant gas power tools; string trimmers, leaf blowers, weed sprayers, etc" – while also addressing labor shortages in the industry – by rolling out emission-free automated helpers.