Discover more from WorkWithAI.com's Daily Newsletter
Leaked info suggests Google Assistant is getting a major AI upgrade
⚡ Today’s Highlights
📰 News: Google Assistant is getting a major AI upgrade + MIT developed a tool that combats unauthorized image manipulation
💰 Funding: Plotlogic, Cyble, Topline Pro, FeelBetter, Grasp, and Rhetorik
🌎 Content: Sal Khan, founder of Khan Academy, discusses how AI is changing education already + a deep dive into C2PA and the fight against the misuse of generative AI
📰 Today's Top Stories
(5 min read) (Source: Axios)
TLDR: Google plans to revamp its Assistant using generative AI technologies like ChatGPT and Bard, aiming to explore a more powerful Assistant experience.
The overhaul will affect how Assistant functions for consumers, developers, and Google's employees, with the company supporting both old and new approaches.
The process has already begun, starting with the mobile version, and involves reorganizing Assistant teams, resulting in a small number of layoffs among thousands of employees working on the Assistant.
The Big Picture: Virtual assistants were one of the first types of AI to permeate modern society so it’s only natural that they’re one of the first systems to get AI upgrades. We are already seeing better AI-assistant integrations with upcoming products like Humane’s AI Pin.
(8 min read) (Source: MIT News)
TLDR: MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) developed "PhotoGuard," an AI tool to combat potential misuse of generative AI models like DALL-E and Midjourney for unauthorized image manipulation.
PhotoGuard creates imperceptible perturbations in pixel values, fortifying images to resist AI-powered manipulation attempts.
The AI employs "encoder" and "diffusion" attacks to confuse an AI model's ability to understand images and manipulates them into producing unrealistic edits, safeguarding images from unauthorized alterations.
The Big Picture: The emergence of advanced generative AI models raises concerns about potential misuse, especially in unauthorized image manipulation, posing risks to privacy, public sentiment, and even financial implications. Many AI models have been trained on content and data without consent from the original owners or creators, causing outrage across industries. As AI technology evolves, ethical considerations and preemptive measures like PhotoGuard are crucial to ensure responsible AI usage and safeguarding against potential misuse in various industries and societal contexts.
(3 min read) (Source: The Verge)
TLDR: Dell is fully embracing the generative AI trend by offering comprehensive solutions to support organizations in harnessing large language models and developing generative AI projects.
They recently announced the introduction of Dell Generative AI Solutions, comprising new hardware setups, a managed service platform, and AI-capable computers for faster AI model processing. Dell partnered with Nvidia, combining Tensor Core GPU with Dell's enterprise AI software and data storage to facilitate efficient AI model execution.
The company also unveiled Dell Professional Services, a platform guiding users through generative AI experimentation and addressing integration challenges within businesses.
(5 min read) (Source: Business Insider)
TLDR: Amazon CEO Andy Jassy has put Rohit Prasad, SVP and head scientist for Alexa, in charge of a new group focusing on the company's "most ambitious" LLMs, which are AI tools that generate human-like responses from massive datasets.
Jassy aims to respond to the growing importance of generative AI, with companies like Microsoft, OpenAI, and Google leading the way, and Meta's Llama 2 model gaining prominence.
Prasad, who has been a key executive for the Alexa team for over a decade, will lead the central team responsible for building Amazon's most expansive LLMs, aligning with Amazon's efforts to make Alexa smarter and more personal using ChatGPT-like technologies.
👀 More Reading:
How Hedge Fund Managers Are Using ChatGPT (Bloomberg)
Can you trust your eyes? (Axios)
How the Pentagon seeds small companies (POLITICO)
The Value of Digital Transformation (Harvard Business Review)
I’m a Hack, by ChatGPT (The New Yorker)
💰 Funding News
The tech provides mining operations a swift and precise geological view, leading to increased yields, operational efficiency and waste reduction
Round co-led by Galvanise Climate Solutions and SE Ventures.
Funding is for driving international expansion (Contracts secured in Indonesia and Brazil, with more potential customers in Chile, Africa, and North America)
Mike Zimmerman, Partner at Main Sequence Ventures, emphasized the significance of mining to accelerate decarbonisation and the global energy transition. “Plotlogic represents a new trend we are seeing, which is not just decarbonisation; it has to be both decarbonisation and productivity gains.”
Their platform uses AI to monitor both dark web and surface web data to map and mitigate companies' digital risk footprints, providing custom threat-hunting solutions for government agencies and Fortune 50 brands.
Round co-led by Blackbird Ventures and King River Capital, with participation from Spider Capital, January Capital, Summit Peak Ventures, and others.
Funding is for general R&D and market expansion.
Zeb Rice, an investor at King River Capital, said, ‘[Its] capable of identifying and guarding against diverse threats from across the web on an unprecedented scale. And the growth and caliber of its customer base stand testament to the global recognition from companies needing best-in-class protection.
Their platform enables small businesses to be easily discovered, trusted, and booked through custom, SEO-optimized websites and applications.
Round led by Forerunner Ventures and supported by Bonfire Ventures, TMV, and BBG Ventures.
Funding is for enhancing the generative AI capabilities, developing new product offerings, and expanding their team.
It identifies high-risk individuals and suggests interventions to reduce hospitalization risks, ensuring safe and effective medication regimens for patients.
Investment led by Firstime Ventures (Keren Kopilov, Partner) and Shoni Health Ventures (Eran Lerer, CEO and Managing Partner), with participation from Random Forest VC (Michael Kerbis, Founding Partner), The Group Ventures, and existing investor Triventures.
Funding is for expanding their reach and refining their solution
5. Grasp, a UK edtech start-up, raised $4M in a Seed funding, for their platform that aims to enhance the rate at which humans learn by providing learners with their optimal online learning experience.
They are starting with modern vocational skills like programming.
Round led by Balderton Capital
Funding is for expanding its product, data science, and engineering teams and further developing the platform
6. Rhetorik, raised $4M CAD for its flagship platform, Neuron360, which empowers businesses with AI-powered insights, actionable intelligence, and targeted lead generation for enterprise marketing and talent sourcing.
Investment led by BDC Capital
Funding is for driving the global expansion of the Neuron360 B2B data platform
🌎 Fresh Content
1. Sal Khan, Founder of Khan Academy, joins Guy Raz on the How I Built This Podcast.
(Spotify: 41 min)
Effects of the pandemic on education and kids.
The email that Sam Altman sent to Sal looking to collaborate.
Khanmigo, the AI personal teaching assistant that Khan Academy is currently developing.
The positive and negative impacts of such a tool on education and on the business itself.
Khan Academy’s vision for the future of education in the age of AI and the role they hope to play.
2. A story we haven’t spent too much time covering, but is of utmost importance in a world of deepfakes and generative AI, is the way copyright will have to adapt and change.
Artists are some of the loudest voices against AI and pressuring institutions through strikes, protests, and lawsuits; The Coalition for Content Provenance and Authenticity (C2PA) was formed as a response to “address the prevalence of misleading information online through the development of technical standards for certifying the source and history (or provenance) of media content.”
bycloud covered C2PA in a deep dive including: (Youtube: 14 min)
A breakdown of the congressional hearings on the topic
The tech that has been developed and proposed as a solution
The goals of C2PA and what the pros and cons of such a solution are
(C2PA is a Joint Development Foundation project, formed through an alliance between Adobe, Arm, Intel, Microsoft and Truepic. )
3. RoomieOfficial went viral a few years back for his "One Guy, 43 Voices" video that garnered over 500M views across the web. He was able to imitate over 43 different artists then, simply with his own voice and a little help from post production.
This time he’s back and imitates over 30 popular artists with complete with deepfakes and AI voice cloning, with this even more impressive song.
(Youtube: 3 min)