THE COMPLETE TEGMARK MULTIVERSE — EXPLAINED! Join us… and find out more!
Subscribe: https://wmojo.com/unveiled-subscribe.
At a conference at New York University in March, philosopher Raphaël Millière of Columbia University offered yet another jaw-dropping example of what LLMs can do. The models had already demonstrated the ability to write computer code, which is impressive but not too surprising because there is so much code out there on the Internet to mimic. Millière went a step further and showed that GPT can execute code, too, however. The philosopher typed in a program to calculate the 83rd number in the Fibonacci sequence. “It’s multistep reasoning of a very high degree,” he says. And the bot nailed it. When Millière asked directly for the 83rd Fibonacci number, however, GPT got it wrong: this suggests the system wasn’t just parroting the Internet. Rather it was performing its own calculations to reach the correct answer.
Although an LLM runs on a computer, it is not itself a computer. It lacks essential computational elements, such as working memory. In a tacit acknowledgement that GPT on its own should not be able to run code, its inventor, the tech company OpenAI, has since introduced a specialized plug-in—a tool ChatGPT can use when answering a query—that allows it to do so. But that plug-in was not used in Millière’s demonstration. Instead he hypothesizes that the machine improvised a memory by harnessing its mechanisms for interpreting words according to their context—a situation similar to how nature repurposes existing capacities for new functions.
This impromptu ability demonstrates that LLMs develop an internal complexity that goes well beyond a shallow statistical analysis. Researchers are finding that these systems seem to achieve genuine understanding of what they have learned. In one study presented last week at the International Conference on Learning Representations (ICLR), doctoral student Kenneth Li of Harvard University and his AI researcher colleagues—Aspen K. Hopkins of the Massachusetts Institute of Technology, David Bau of Northeastern University, and Fernanda Viégas, Hanspeter Pfister and Martin Wattenberg, all at Harvard—spun up their own smaller copy of the GPT neural network so they could study its inner workings. They trained it on millions of matches of the board game Othello by feeding in long sequences of moves in text form. Their model became a nearly perfect player.
Proponents of Direct Fusion Drive (DFD) propulsion system say the drive could send us to Saturn’s moon, Titan, in just 2 years. Here’s how it works.
‘Star Trek: Picard’ is set 400 years in the future, but, like most science fiction, it deals with issues in the here and now. The show’s third and final season provides a lens on cybersecurity.
This is awesome and more than a little scary. You can actually try it on a youtube video and then ask questions about the content…
Introducing LeMUR, AssemblyAI’s new framework for applying powerful LLMs to transcribed speech.
With a single line of code, LeMUR can quickly process audio transcripts for up to 10 hours worth of audio content, which effectively translates into ~150k tokens, for tasks likes summarization and question answer.
Continue reading “SpeechGPT? LeMUR AI can turn 10 Hrs SPEECH into LLM Context” »
High spatial and temporal resolution
The scientists performed all these analyses on organoids that were of different ages and thus at different stages of development. In this way, they were able to create a time series of images and genetic information that describes the entire 39-week development of retinal organoids.
A new way to gauge the universe’s expansion rate has delivered a confusing result that may conflict with previous related measurements.
The content of this post is solely the responsibility of the author. AT&T does not adopt or endorse any of the views, positions, or information provided by the author in this article.
Intro
In February 2022, Microsoft disabled VBA macros on documents due to their frequent use as a malware distribution method. This move prompted malware authors to seek out new ways to distribute their payloads, resulting in an increase in the use of other infection vectors, such as password-encrypted zip files and ISO files.
The ability to build AI systems and the idea empower a business to scale them is on the rise. Read about how AI could empower any business at Landing AI.
Summary: Researchers have created a device that emulates the human eye’s ability to see color by using narrowband perovskite photodetectors and a neuromorphic algorithm.
The photodetectors, sensitive to red, green, and blue light, mimic our cone cells, while the neuromorphic algorithm simulates our neural network to process information into high-quality images. Unlike modern cameras that require external filters, this technology could improve resolution and reduce manufacturing costs.
The device also generates electricity as it absorbs light, potentially leading to battery-free camera technology.
