- The Tech Buzz
- Posts
- 9/22/23
9/22/23
∆ Tech in the News
"Light-Touch" AI Bill - Senate Minority Whip John Thune (R-S.D.) and Sen. Amy Klobuchar (D-Minn) plan to introduce a bill requiring AI companies to self-certify the safety of high-risk AI systems.
Game of Drones - Literary titans including George R.R. Martin, and John Grisham join an Authors’ Guild suit vs. OpenAI over ChatGPT allegedly condensing their works without due compensation or permission.
Alexa Becomes “Super Agent” - Amazon's Fall Product Event highlighted upgraded gadgets, faster AI and more smart home integration, with Alexa gaining “more conversational” AI skills.
Ask The Oracle - Oracle just filed a patent app for “human-understandable insights” to help AI model users understand how they arrived at their answers.
“No Cloud Needed” - Intel unveiled and demoed its 'AI PC' - a new chip to run ChatGPT-like tech privately on local devices.
Is a "light Touch" self-certification AI regulation strong enough? |
˚The hottest thing in tech this week
OpenAI and Google Race for Multimodal, Interactive AI
As Google’s Gemini launch approaches, OpenAI is hustling to beat them to the drop with the next generation of ‘multimodal’ and ‘interactive’ AI. No better sign that OpenAI sees Google’s soon-to-drop Gemini model as a serious threat.
So, What is Multimodal AI? Simply, it’s AI that can process, understand, and generate outputs for more than one type of data ‘modality’. Interactive AI, meanwhile is AI that can carry out tasks by calling on other software and other people to get stuff done.
OpenAI’s Future
OpenAI just released DALL-E 3 the latest version of its AI image generator, soon to be integrated with ChatGPT as "GPT-Vision".
The new "Gobi," model is reportedly designed as multimodal from the ground up and could eventually become GPT-5.
Microsoft just integrated OpenAI GenAI features into its “Co-pilot” for Windows and the Office 365 suit for Word, PowerPoint and Excel.
Google’s Gemini Vision
Deepmind’s new Gemini model aims to compete with ChatGPT 4.
CEO Sundar Pichai has said Gemini is designed from the ground up to be multimodal, integrating text, images, and other data types.
Pichai also hinted at it being an “incredible universal personal assistant” integrated throughout people’s daily travel, work, and entertainment.
The new Bard Extensions, for example, already pull from Google tools and services like Gmail, Calendar and Drive. Gemini will upgrade this.
It’s hard to predict a winner, but it’s clear the new approaches by these giants are defining the future direction for AI development as we move beyond the GenAI wave to more conversational, interconnected and empowered AI agents.
Who do you back to own the multimodal and interactive future?
“Gemini will be multimodal from the ground up. We’ve done a lot of work in the past that has underpinned multimodal work across the industry. Gemini is several improvements on top of that.”
News Bytes
FDA Gives Nod to Neuralink Human Trials for Paralysis
AI Data Analysis Helping Battle Wildfires in Alberta, Canada
Chan Zuckerberg Initiative Bets on AI for Disease Breakthroughs
Deals We’re Watching
Founders to Watch and Opportunities to Catch
Closed Deals:
Building “Knowledge Fabric” - Pryon Sealed a $100M round led by the U.S. Innovative Tech Fund to offer AI-driven enterprise data processing and indexing.
Pivot to Software/AI - Cisco just snapped up analytics and cybersecurity firm Splunk for $28B. Cisco-Splunk will be one of the world's largest software firms, gaining strong ARR and powerful AI initiatives and capacity.
Roblox Buys AI Chat Startup - The buyout of voice API startup Speechly will bring its advanced real-time voice chat and AI Voice Chat Moderation services to the online gaming network.
Gut Feeling - Vivante Health built a digital platform for managing chronic digestive conditions. A recent $31M Series B led by Mercato Partners attracted backing from Health Catalyst Capital.
AI-Driven Clinical Research - Healthtech startup Medeloop raised an $8M Seed to advance its AI-powered data platform and partner with pharma companies on enhanced clinical trials.
Companies to Watch:
Open-Source Analytics Processing - MotherDuck was founded by an ex-Google engineer, backed by a16z, now has a live public analytics database, and just closed a $52.5M Series B at a $400M post-money valuation.
Navigating The AI Market - Mapbox closed a $280M Series E led by SoftBank for their real-time mapping AI that powers navigation for Toyota, General Motors and BMW.
Web 3 “Connection Layer” - Mesh is a startup for digital asset holders to manage assets across blockchains and platforms with 70 paying industry clients. A $22M raise was backed by Money Forward, Galaxy, and Samsung.
AI Space Planning - Qbiq offers genAI-powered space planning and visualization tools. Their $10M Seed Round just closed with backing from JLL Spark Global Ventures, 10D, Ocean Azul, Randomforest and M-FUND.
Shaq Leads Round - The NBA All-Star led a Seed round for AI reading education app Edsoma raising $2.5M from 12+ individual investors at a post-investment valuation of $14M. The app is live on Play and the App Store now.
Open Deals:
Tackling Charge Anxiety - EVLUV brings an “OpenTable-inspired” EV charger solution to the market. Currently raising a $500K Angel Round. Schedule a chat with CEO & Co-Founder Guzel Lumpkin here to learn more.
“Easy as Calling a Friend” - Teleperson uses AI to redesign customer service by directing consumers to the most efficient, or preferred path (chatbot or human rep), avoiding “phone jail or call trees.” Schedule a call here.
IPOs Are Back:
Is It “Time to IPO”? - Instacart’s highly anticipated IPO raised $660M, and its founder exited with $1.1B. But, the IPO also slashed 75% from the last private valuation and left some late investors underwater.
Klaviyo’s Big Day - The Marketing automation firm sold 19.2M shares at a market value of $9.2B on it IPO off the back of 50% YoY revenue growth. Shares popped 22.5% on the first day of trading post-IPO but retraced these gains.
Tools
Debunkd - helps you fact-check any ChatGPT result
1PhotoAI - generates countless headshots from 1 selfie using AI
DALL-E 3 - the latest version of its Open AI’s image generator
If you have a deal or company that you would like to host on our next issue, drop us a line at @ [email protected].
How did we do this time? |
Thanks for connecting with us.
The Tech Buzz Team