Friday, January 31, 2025
HomeRoboticsThese Had been the ten Most Well-liked SingularityHub Tales of 2024

These Had been the ten Most Well-liked SingularityHub Tales of 2024


This yr, readers had been once more fascinated by tales about synthetic intelligence. One algorithm realized to make brief, playable video video games from video footage; one other cloned actual folks’s personalities; and yet one more took on the position of your future self—prepared and keen to impart knowledge.

Different well-liked items dug into the rise of robotaxis, tantalizing hints about how we would in the future stave off getting old, and a unified concept of consciousness.

We hope you get pleasure from a re-assessment or discovering these for the primary time.

As all the time, thanks for studying!

google deepmind generative ai video games and world models
Credit score: Google DeepMind

A Google AI Watched 30,000 Hours of Video Video games—Now It Makes Its Personal
By Jason Dorrier
“As AI requires prodigious quantities of knowledge, one option to forecast the place issues are going subsequent is to take a look at what knowledge is extensively out there on-line, however nonetheless largely untapped. Video, of which there’s lots, is an apparent subsequent step. Certainly, [in February], OpenAI previewed a brand new text-to-video AI known as Sora that shocked onlookers. However what about video…video games?”

Scientists Are Working In the direction of a Unified Idea of Consciousness
By Shelly Fan
“Not all [the] authors agree on the precise mind mechanisms that permit us to understand the outer world and assemble an inside world of ‘self.’ However by collaborating, they merged their concepts, displaying that completely different theories aren’t essentially mutually incompatible—actually, they might be consolidated right into a common framework of consciousness and even encourage new concepts that assist unravel one of many mind’s biggest mysteries.”

Scientists Lengthen Life Span in Mice by Restoring This Mind-Physique Connection
By Shelly Fan
“Altering the protein’s habits in aged mice with genetic engineering prolonged their life span by roughly seven p.c. For a median 76-year life span in people, the rise interprets to over 5 years. The therapy additionally altered the mice’s well being. Mice like to run, however their vigor plummets with age. Reactivating the neurons in aged mice revived their motivation, reworking them from sofa potatoes into spectacular joggers.”

These were our favorite science and tech stories from around the web this week.
Credit score: SpaceX

It Will Take Solely a Single SpaceX Starship to Launch a House Station
By Edd Gent
“SpaceX’s forthcoming Starship rocket will make it doable to elevate unprecedented quantities of fabric into orbit. …Now, a three way partnership between Airbus and Voyager House that’s constructing a non-public area station known as Starlab has inked a contract with SpaceX to get it into orbit. The enterprise plans to place the spectacular capabilities of the brand new rocket to full use by launching your entire 26-foot-diameter area station in a single go.”

OpenAI’s GPT-4o Makes AI Clones of Actual Folks With Shocking Ease
By Edd Gent
“AI has change into uncannily good at aping human conversational capabilities. New analysis suggests its powers of mimicry go lots additional, making it doable to duplicate particular folks’s personalities. …A examine led by researchers at Stanford College has found that every one it takes is a two-hour interview for an AI mannequin to foretell folks’s responses to a battery of questionnaires, character checks, and thought experiments with an accuracy of 85 p.c.”

‘Droidspeak’: AI Brokers Now Have Their Personal Language Because of Microsoft
By Edd Gent
“Getting AIs to work collectively might be a robust drive multiplier for the know-how. However regardless of their expressive energy, human languages may not be the very best medium of communication for machines that basically function in ones and zeros. This prompted researchers from Microsoft to develop a brand new methodology of communication that permits brokers to speak to one another within the high-dimensional mathematical language underpinning LLMs. They’ve named the brand new method Droidspeak…and in a preprint paper revealed on the arXiv, the Microsoft workforce experiences it enabled fashions to speak 2.78 instances sooner with little accuracy misplaced.”

Microsoft created a new machine language for AI agents communicating with each other.
Credit score: Shawn Suttle from Pixabay

Waymo Robotaxis Are Giving 100,000 Rides a Week. It’ll Quickly Be Extra.
By Jason Dorrier
“After a yr of economic operations in San Francisco with out main incident, Waymo is eyeing growth. In August, the corporate moved into Daly Metropolis, Broadmoor, and Colma, simply south of town. Waymo has approval to function in a complete of twenty-two cities alongside the peninsula south of San Francisco, and though there’s no timetable but, in keeping with the San Francisco Chronicle, additionally they have ambitions so as to add operations in San Jose and East Bay, which would come with Oakland and Berkeley.”

A One-and-Completed Injection to Sluggish Ageing? New Research in Mice Opens the Chance
By Shelly Fan
“A preventative anti-aging remedy looks like wishful considering. But a brand new examine led by Dr. Corina Amor Vegas at Chilly Spring Harbor Laboratory describes a therapy that brings the dream to life—at the very least for mice. Given a single injection in younger maturity, they aged extra slowly in comparison with their friends. …’If we give it to aged mice, they rejuvenate. If we give it to younger mice, they age slower. No different remedy proper now can do that,’ stated Amor Vegas in a press launch.”

These Mini AI Fashions Match OpenAI With 1,000 Instances Much less Knowledge
By Jason Dorrier
“The factitious intelligence trade is obsessive about dimension. Greater algorithms. Extra knowledge. Sprawling knowledge facilities that would, in a number of years, eat sufficient electrical energy to energy entire cities. …Eye-popping numbers like these make it simple to neglect dimension isn’t all the things. Some researchers, notably these with fewer sources, are aiming to do extra with much less. AI scaling will proceed, however algorithms can even get way more environment friendly as they develop.”

This MIT Chatbot Simulates Your ‘Future Self.’ It’s Right here to Assist You Make Higher Selections.
By Jason Dorrier
“Chatbots at the moment are posing as mates, romantic companions, and departed family members. Now, we are able to add one other to the listing: Your future self. MIT Media Lab’s Future You challenge invited younger folks, aged 18 to 30, to have a chat with AI simulations of themselves at 60. The sims—which had been powered by a personalised chatbot and included an AI-generated picture of their older selves—answered questions on their expertise, shared recollections, and supplied classes realized over the a long time.”

Banner Picture Credit score: Luke Jones on Unsplash

RELATED ARTICLES

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Most Popular

Recent Comments