Google Cloud VP says GenAI is a solution to data analytics
August 9, 2024
Good morning, and happy Friday. Yesterday, OpenAI gained a new board member in the form of Carnegie Melon machine learning professor Zico Kolter. He will additionally be joining the company’s safety team.
In other news, we chatted with a Google Cloud VP about generative AI and data analytics, specifically what this all means for the people working in the field.
Read on for the full story.
— Ian Krietzberg, Editor-in-Chief, The Deep View
In today’s newsletter:
AI for Good: Fighting invasive plant species
Source: BirdHabitatBot
Just as Florida is dealing with an invasion of non-native lizards, Connecticut is dealing with an invasion of non-native plants.
Two of the plants in question — Japanese barberry and multiflora rose — are a constant problem in Connecticut forests; they choke out native flora, conquering entire areas and threatening forest ecosystems.
Current methods of removal involve either a propane torch or chemical herbicide, both of which pose varying levels of risk to the environment. One researcher — Nancy Marek — sought a better way to address these invasive plants. She found an answer in AI, drones and robotics.
The details: Marek’s project — named BirdHabitatBot — involves three stages.
She starts by gathering drone images of a forest; she then runs a deep learning AI model over the footage, which was trained to identify specific instances of Japanese barberry and multiflora rose.
This alone allows conservationists to take a more efficient approach toward addressing these invasive species, but Marek is also working on developing a robot that, armed with the AI algorithm’s information, would be able to march into the undergrowth and remove the offending plants.
“You can’t fix all the problems,” Marek said. “But if this is one problem I can at least make a dent in, that would be glorious.”
Want to close your books in record time and get back to growing your business? BILL Spend & Expense's free spend & expense management software helps you:
Fight fraud with enforceable budgets and unlimited virtual cards1
Maximize rewards from ad spend and subscriptions
Save time on expense reporting and closing the books
Access credit lines from $500-5M2 as you scale
Antitrust regulators are investigating Amazon’s partnership with Anthropic
Source: Anthropic
Britain’s competition watchdog on Thursday said it had garnered “sufficient information” to investigate Amazon’s partnership with Anthropic.
The organization now has until Oct. 4 to decide whether to clear the partnership of scrutiny or deepen the probe with a “Phase 2” investigation.
The context: Amazon has invested $4 billion in Anthropic; $1.25 billion in September and another $2.75 billion earlier this year.
As part of the partnership, Amazon has made Anthropic’s models available on its Bedrock platform, and Anthropic has built its models using Amazon’s chips.
The same competition authority is also investigating Microsoft’s partnership with Inflection and recently launched a probe into Google’s relationship with Anthropic.
“We’re disappointed that the UK’s Competition and Markets Authority (CMA) has not ended its probe yet,” an Amazon spokesperson told me. “Amazon’s collaboration with Anthropic does not raise any competition concerns or meet the CMA’s own threshold for review.”
They added that model-building is expensive and “companies like Anthropic need access to a substantial amount of capital to train these models.” The spokesperson said that “Amazon holds no board seat nor decision-making power at Anthropic,” adding that it will continue to make Anthropic’s models available on Bedrock.
Anthropic did not respond to a request for comment.
Database startup Hyperspace recently closed a $9.5 million seed funding round.
Heeyo, an AI app for kids, recently raised $3.5 million in seed funding.
Google and Meta struck secret ads deal to target teenagers (FT).
Exclusive: Renowned Experts Pen Support for California’s Landmark AI Safety Bill (Time).
Hugging Face makes its largest-ever acquisition so it can host more models (Forbes).
Amazon Enlists TikTok, Pinterest in Quest to Sell Everywhere (The Information).
Palantir jumps 11% on Microsoft partnership to sell AI to U.S. defense, intel agencies (CNBC).
Watch this robot win at Ping Pong
Source: Unsplash
A team at Google DeepMind has trained a robot to play Table Tennis. Of the 29 matches it played, the robot wound up winning 45% of them. Against beginners, the robot won all matches; against advanced players, the robot lost all matches, landing the robot’s skill at somewhere in the “intermediate” gradient (I don’t think it could beat me, but it might be close).
Why’d they do it: Since the sport requires physical skill, speed and accuracy, in addition to real-time and strategic decision-making, the team said it’s a “valuable benchmark for advancing robotic capabilities.”
The team collected a small amount of human-human play data to start with, then trained the system in simulation before setting it against human players. It consists of a “library of low-level skills (backhand, forehand, etc) and a high-level controller that selects the most effective skill.”
The system continuously generates more training data as it plays.
Advanced players saw the potential for the robot to act as a “more dynamic practice partner.”
Invest smarter with Public and AI
If you're looking to leverage AI in your investment strategy, you need to check out Public.
The all-in-one investing platform allows you to build a portfolio of stocks, options, bonds, crypto and more, all while incorporating the latest AI technology — for high-powered performance analysis — to help you achieve your investment goals.
Join Public, and build your primary portfolio with AI-powered insights and analysis.
Google Cloud VP says genAI is a solution to data analysis
Source: Unsplash
We live in a data-driven world. And, unsurprisingly, there exists an entire field whose explicit purpose is to help everyone else make sense of all that data. That field, of course, is data science.
And despite the volatility of the tech sector — specifically amid the rise of generative AI — data scientists remain as valuable as ever.
365 Data Science’s 2023 tech layoffs study found that data scientists made up only 3% of cut positions, whereas software engineers made up 22%.
As 365 noted, the rise of generative AI — which is really good at deriving insights from giant swaths of data, the key description of a data scientist’s job — has, far from making the field obsolete, highlighted the “indispensable nature of data science skills.”
I sat down with Google Cloud’s VP/GM for data analytics Gerrit Kazmaier to talk about the role of generative AI in data analysis, and what impact this is having on the field.
The big way that generative AI has changed and progressed the field, according to Kazmaier, is that it allows companies to work with unstructured data (documents, pictures, videos) with the same kind of flexibility as they would work with structured data.
He said that the incorporation of genAI allows companies to scale up their scarcest resource: data scientists, analysts and engineers.
“There are not many people who have a high proficiency in working with data and answering questions based off of data,” Kazmaier said. “And that's a key constraint virtually any company is facing.”
Enhancement, not replacement: I hear often that genAI will be a tool to enhance professionals, not replace them. The famous iteration of this, of course, is the rather annoying phrase: “you won’t be replaced by AI, you’ll be replaced by someone who uses AI.”
In the case of data science, according to Kazmaier, it’s not a question of replacement since there aren’t enough data scientists out there to respond to the data science needs of the world.
“It’s a major enhancement,” he said. “Companies have much more data and much more scenarios for data — many more questions of their data — than what they can actually find, hire and train for in these roles.”
He said that Google’s AI-enabled data platform BigQuery has 17 capabilities specifically designed to make data scientists work more quickly and more efficiently.
“I think data analysts and data engineers and data scientists, a lot of their skills is not only in generating prompts; a lot of their skill is actually asking the right questions,” Kazmaier said. “Reasoning deeply about the business and the data and coming up with genuine insights.”
Automation “frees up time to ask way more interesting questions.”
Which image is real? |
A poll before you go
Thanks for reading today’s edition of The Deep View!
We’ll see you in the next one.
Your view on California SB 1047:
Roughly half of you said the bill is either better than nothing or really good.
Around 20% of you said you hate it and another 20% said it’s okay but it needs work.
It’s okay, but needs work:
“Let's find accurate and dependable ways to stay safe.”
Something else:
“It’s too early for such broad regulation. For now, we need to focus on more narrow regulation, targeting clearly harmful areas like kids creating fake nudes of their classmates. We need more experience and a broader understanding of the impact of AI before taking broad regulatory brush strokes that slow down AI’s evolution.”
Would you play a robot in Table Tennis? |
*Public disclosure: All investing involves the risk of loss, including loss of principal. Brokerage services for US-listed, registered securities, options and bonds in a self-directed account are offered by Public Investing, Inc., member FINRA & SIPC. Cryptocurrency trading services are offered by Bakkt Crypto Solutions, LLC (NMLS ID 1828849), which is licensed to engage in virtual currency business activity by the NYSDFS. Cryptocurrency is highly speculative, involves a high degree of risk, and has the potential for loss of the entire amount of an investment. Cryptocurrency holdings are not protected by the FDIC or SIPC.
Alpha is an experiment brought to you by Public Holdings, Inc. (“Public”). Alpha is an AI research tool powered by GPT-4, a generative large language model. Alpha is experimental technology and may give inaccurate or inappropriate responses. Output from Alpha should not be construed as investment research or recommendations, and should not serve as the basis for any investment decision. All Alpha output is provided “as is.” Public makes no representations or warranties with respect to the accuracy, completeness, quality, timeliness, or any other characteristic of such output. Your use of Alpha output is at your sole risk. Please independently evaluate and verify the accuracy of any such output for your own use case.