Figure announces $70M in series A funding. The company is developing a multi-purpose humanoid robot that could one day take on a wide range of tasks traditionally performed by people.
The robots are coming.
The robots are coming.
TechCrunch
Figure raises $70M to build its humanoid robots
I spent a bit of time at Figure’s Sunnyvale offices during a recent visit to the South Bay. The firm is currently in that semi-awkward phase where it’s ready to talk about what it’s working on — but only to a point. That means things like its process and…
The False Promise of Imitating Proprietary LLMs
Open-sourced LLMs are adept at mimicking ChatGPT’s style but not its factuality. There exists a substantial capabilities gap, which requires better base LM.
If you want GPT-4 performance you should use way more than 4 open source language models working together.
One to one you’ll never get the same factuality in particular.
It requires new design patterns as well as better base LMs which will take time & effort
Open-sourced LLMs are adept at mimicking ChatGPT’s style but not its factuality. There exists a substantial capabilities gap, which requires better base LM.
If you want GPT-4 performance you should use way more than 4 open source language models working together.
One to one you’ll never get the same factuality in particular.
It requires new design patterns as well as better base LMs which will take time & effort
Great work from TII in the UAE releasing open source 7b and 40b models as well as a new high quality dataset.
7b https://huggingface.co/tiiuae/falcon-7b
40b https://huggingface.co/tiiuae/falcon-40b
7b https://huggingface.co/tiiuae/falcon-7b
40b https://huggingface.co/tiiuae/falcon-40b
huggingface.co
tiiuae/falcon-7b · Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
👍1
The Grace Hopper Superchip architecture, shown off by Nvidia's Jensen Huang is super geeky but fascinating.
It basically marries the specialization of the Hopper GPU with the generalization of the Grace CPU by allowing access to all memory.
Here we have a more classic architecture.
Remember that memory cannot be accessed if it cannot be addressed done through a page table. In this traditional approach, GPU & CPU run different page tables.
With a GH Superchip, one page table exists & translates memory addresses so GPU can access CPU memory (here it's LPDDR5X) and CPU can read the GPU's memory (HBM3).
Curious to know how devs will utilize this approach.
It basically marries the specialization of the Hopper GPU with the generalization of the Grace CPU by allowing access to all memory.
Here we have a more classic architecture.
Remember that memory cannot be accessed if it cannot be addressed done through a page table. In this traditional approach, GPU & CPU run different page tables.
With a GH Superchip, one page table exists & translates memory addresses so GPU can access CPU memory (here it's LPDDR5X) and CPU can read the GPU's memory (HBM3).
Curious to know how devs will utilize this approach.
BiomedGPT: a Unified and Generalist Biomedical Generative Pre-trained Transformer for Vision, Language, and Multimodal Tasks
Outperforms the majority of preceding SotA models across 5 tasks with 20 datasets spanning over 15 biomedical modalities.
Outperforms the majority of preceding SotA models across 5 tasks with 20 datasets spanning over 15 biomedical modalities.
RIP online job interviews.
This AI tool enables real-time transcriptions for your microphone input AND speaker output.
It then generates a response for the user to answer questions based on the live conversation:
The tool uses Whisper for transcriptions and GPT-3.5 for suggested responses.
This AI tool enables real-time transcriptions for your microphone input AND speaker output.
It then generates a response for the user to answer questions based on the live conversation:
The tool uses Whisper for transcriptions and GPT-3.5 for suggested responses.
GitHub
GitHub - SevaSk/ecoute: Ecoute is a live transcription tool that provides real-time transcripts for both the user's microphone…
Ecoute is a live transcription tool that provides real-time transcripts for both the user's microphone input (You) and the user's speakers output (Speaker) in a textbox. - SevaSk/ecoute
🔥1😁1
New paper: “The Economics of Augmented and Virtual Reality”
The focus is on what provides value — particularly to decision-making. This allows to explicitly rule out what will be of low value — that turns out to be much of the focus of AR/VR up until now.
For AR, value is created when there is high contextual entropy (that is, there is a ton of information in the environment) and the tech allows you to reduce that information and the cognitive load of processing it.
Thus, AR that involves popping up notifications in your eye-line are not of high value — they increase cognitive load of the environment you are in. So Google Glass, North etc were on the wrong path.
For VR, providing nice looking environments for meetings with people you already know provides less information than a zoom call. Thus, Meta-style virtual meetings are not worth the effort for most things. They have to be better than zoom and it is unclear VR will achieve that.
The focus is on what provides value — particularly to decision-making. This allows to explicitly rule out what will be of low value — that turns out to be much of the focus of AR/VR up until now.
For AR, value is created when there is high contextual entropy (that is, there is a ton of information in the environment) and the tech allows you to reduce that information and the cognitive load of processing it.
Thus, AR that involves popping up notifications in your eye-line are not of high value — they increase cognitive load of the environment you are in. So Google Glass, North etc were on the wrong path.
For VR, providing nice looking environments for meetings with people you already know provides less information than a zoom call. Thus, Meta-style virtual meetings are not worth the effort for most things. They have to be better than zoom and it is unclear VR will achieve that.
👍2❤1💩1
MindEye is a state-of-the-art fMRI-to-image approach that retrieves and reconstructs images from brain activity
Project page: medarc-ai.github.io/mindeye/
arXiv: arxiv.org/abs/2305.18274
Project page: medarc-ai.github.io/mindeye/
arXiv: arxiv.org/abs/2305.18274
🔥1
Qualcomm’s working on software that will extend phone apps to AR headsets. It’s a hint of where Google and Apple both may be heading.
CNET
Qualcomm's Phone-Connected AR Headset Plans Hint at Google and Apple's Future
Future headsets will need to work better with phone apps: Qualcomm's moves point to where Google and Apple need to go.
Google invests in AI startup Runway to wrest cloud business from AWS
Google is investing in Runway a startup that lets customers generate video from text descriptions using artificial intelligence it pioneered, at a valuation of around $1.5 billion including the new capital.
The investment, part of a roughly $100 million funding round, underscores the fierce competition among cloud providers to get close to companies with cutting-edge AI services that could become major cloud customers or acquisition targets in the future.
Amazon Web Services has touted Runway, which generates relatively little revenue from its video-editing tools, as a key AI-startup customer but Runway is now expected to rent cloud servers from Google.
Google is investing in Runway a startup that lets customers generate video from text descriptions using artificial intelligence it pioneered, at a valuation of around $1.5 billion including the new capital.
The investment, part of a roughly $100 million funding round, underscores the fierce competition among cloud providers to get close to companies with cutting-edge AI services that could become major cloud customers or acquisition targets in the future.
Amazon Web Services has touted Runway, which generates relatively little revenue from its video-editing tools, as a key AI-startup customer but Runway is now expected to rent cloud servers from Google.
The Information
Google Invests in AI Startup Runway to Wrest Cloud Business From AWS
Google is investing in Runway, a New York–based startup that lets customers generate video from text descriptions using artificial intelligence it pioneered, at a valuation of around $1.5 billion including the new capital, according to two people familiar…
Interesting that Tesla is hiring machine learning engineers to work on large language models:
"Machine Learning Engineer, LLM, Autopilot AI (Dojo)"
What does this mean for the future of their FSD?
"Machine Learning Engineer, LLM, Autopilot AI (Dojo)"
What does this mean for the future of their FSD?
Tesla Careers
Machine Learning Engineer, LLM, Autopilot AI (Dojo) | Tesla Careers
Apply for the Machine Learning Engineer, LLM, Autopilot AI (Dojo) position in PALO ALTO, California.
Japan recently reaffirmed that it will not enforce copyrights on data used in AI training
The policy allows AI to use any data “regardless of whether it is for non-profit or commercial purposes, whether it is an act other than reproduction, or whether it is content obtained from illegal sites or otherwise.”
The policy allows AI to use any data “regardless of whether it is for non-profit or commercial purposes, whether it is an act other than reproduction, or whether it is content obtained from illegal sites or otherwise.”
ReachNeuro announced breakthrough status for its spinal cord stimulation device to treat paralysis from stroke.
The company is a spin-out from a University of Pittsburgh and Carnegie Mellon lab, which earlier this year restored hand movement for two stroke survivors.
The company is a spin-out from a University of Pittsburgh and Carnegie Mellon lab, which earlier this year restored hand movement for two stroke survivors.
Zuckerberg teased his company's latest virtual reality headset the Quest 3
The new headset will start at $499, Zuckerberg said in a video on Instagram. The post noted that it'll be the first announced headset with "high-res color mixed reality." The Quest 3 is 40% thinner than its predecessor, the Quest 2.
It appears to have at least three cameras on the front that may improve some of the pass-through experiences, where a user can see the real world around them while wearing the headset and interacting with apps.
The video noted that it had improved "natural" depth perception for the wearer, and touted graphics improvements for gaming.
"High-fidelity color Passthrough, innovative machine learning, and spatial understanding let you interact with virtual content and the physical world simultaneously, creating limitless possibilities to explore," Meta said in a release.
The Quest 3 will ship with a next-generation Qualcomm chipset, the post said. Meta said it will provide more details on Sept. 27 during its Connect conference.
Meta's Quest 2 headset was released in the fall of 2020 at a starting price of $299.
The new headset will start at $499, Zuckerberg said in a video on Instagram. The post noted that it'll be the first announced headset with "high-res color mixed reality." The Quest 3 is 40% thinner than its predecessor, the Quest 2.
It appears to have at least three cameras on the front that may improve some of the pass-through experiences, where a user can see the real world around them while wearing the headset and interacting with apps.
The video noted that it had improved "natural" depth perception for the wearer, and touted graphics improvements for gaming.
"High-fidelity color Passthrough, innovative machine learning, and spatial understanding let you interact with virtual content and the physical world simultaneously, creating limitless possibilities to explore," Meta said in a release.
The Quest 3 will ship with a next-generation Qualcomm chipset, the post said. Meta said it will provide more details on Sept. 27 during its Connect conference.
Meta's Quest 2 headset was released in the fall of 2020 at a starting price of $299.
CNBC
Zuckerberg unveils Meta Quest 3 VR headset days before Apple reveals its own
The Meta Quest 3 is the successor to the Quest 2 headset that was released in the fall of 2020 at a starting price of $299.
Deloitte_Metaverse_Report_May_2023_1685639206.pdf
15.1 MB
Deloitte introduced an interesting report on the Metaverse
According to the report, the Metaverse could contribute US$560 billion per year to US GDP by 2031.
This report includes:
- What is the Metaverse
- The opportunity for the US economy
- How (and what) to build in the Metaverse
- Helpful case studies
- Glossary of Metaverse terms
According to the report, the Metaverse could contribute US$560 billion per year to US GDP by 2031.
This report includes:
- What is the Metaverse
- The opportunity for the US economy
- How (and what) to build in the Metaverse
- Helpful case studies
- Glossary of Metaverse terms
🔥3❤1
With a new AI algorithm, finally managed to discover new conservation laws that domain experts didn't know about, in both fluid mechanics and atmospheric chemistry.
🆒1
Galaxy_2023_Watch_This_Space_1685699522.pdf
16.4 MB
Terrific Q1 overview of what’s happening in crypto & Web3 by Galaxy
A few highlights:
– Year to date, Bitcoin is the best performing asset when compared to a range of securities (both equities and fixed income), indices, and commodities
– Bitcoin's hashrate grew almost 30% in Q1 2023 to more than 340 EH/sec.
– Ethereum continues to dominate user activity and developer mindshare
– The leading optimistic rollups, Optimism and Arbitrum, experienced a 10x increase in daily active addresses since the start of 2022
– NFTs on Bitcoin could reach a market cap of $4.5bn in two years
On Web3
“Builders can make use of open architecture to drive greater innovation. Creators and companies can interact with customers and monetize intellectual property in new and enhanced ways, and users can regain and retain control over their personal data and participate more holistically in the platforms they use.”
A few highlights:
– Year to date, Bitcoin is the best performing asset when compared to a range of securities (both equities and fixed income), indices, and commodities
– Bitcoin's hashrate grew almost 30% in Q1 2023 to more than 340 EH/sec.
– Ethereum continues to dominate user activity and developer mindshare
– The leading optimistic rollups, Optimism and Arbitrum, experienced a 10x increase in daily active addresses since the start of 2022
– NFTs on Bitcoin could reach a market cap of $4.5bn in two years
On Web3
“Builders can make use of open architecture to drive greater innovation. Creators and companies can interact with customers and monetize intellectual property in new and enhanced ways, and users can regain and retain control over their personal data and participate more holistically in the platforms they use.”