RECEIVE THE DAILY AI 3 TIMES PER WEEK TO YOUR INBOX. SUBSCRIBE NOW
This week, geopolitical industry tensions took a backseat in favor of increased competition and innovation.
Additionally, tech firms continued to churn out LLMs with increased bias on human-oriented data.
Here are the headlines:
Google revealed the launch of its Gemma LLM;
Open AI released CriticGPT, an LLM that checks ChatGPT;
AWS is investigating web scraping allegations (against Perplexity AI);
LLMs can (now) pass examinations better than human students;
Plus, Bill Gates allayed AI energy consumption fears;
And much more!
Of course, it's your Daily AI™!
The DWN AI Index™ Update
On 6/27/24, the index remained (slightly) positive with a +8.86-point change, at +0.28%, to close at 3164.25.
DWN/AI Index™ Individual Stock Closing Prices:
Amazon (AMZN) at $197.85,
Arista Networks (ANET) at $347.93,
C3.ai (AI) at $28.55,
CrowdStrike Holdings (CRWD) at $387.29,
Duolingo (DUOL) at $208.65,
iRhythm Technologies (IRTC) at $104.38,
Microsoft Corporation (MSFT) at $452.85,
NVIDIA Corporation (NVDA) at $123.99,
Palantir Technologies (PLTR) at $25.22,
Taiwan Semiconductor Manufacturing (TSM) at $171.63,
Meta Platforms, Inc. (META) closed at $519.56, while Alphabet Inc. (GOOG) closed at $186.86.
Under the AI Hood
Meta
According to a June 27 TechCrunch report, social media giant Meta CEO Mark Zuckerberg revealed the (initial) testing of user-created AI models on Instagram. Zuckerberg (also) iterated that the chatbots will have AI watermarks. Additionally, testing starts in the United States.
In a June 27 blog post, Google introduced Gemma 2. The latest version of its open-source LLM has 9 billion (9B) and 27 billion (27B) parameter sizes and outperforms several LLMs which are twice as large. Google also affirmed its commitment to responsible (and ethical) AI deployments, revealing the availability of its Responsible Generative AI Toolkit.
Furthermore, Gemma 2's model weights are (currently) available for researchers and developers via Hugging Face and Kaggle. The search engine behemoth (also) unveiled support for an (additional) 110 new languages via the deployment of its PaLM 2 (language) model.
Open AI
Interestingly, Open AI revealed in a June 27 research paper the launch of CriticGPT, an LLM that reviews responses from its (popular) ChatGPT LLM series. CriticGPT is based on ChatGPT-4 and serves as an AI-based peer review solution to (help) detect errors in generated output.
Open AI also asserted a 60% improvement in accuracy when users deployed CriticGPT. Consequently, the tech firm unveiled plans to integrate similar models in reinforcement learning from human feedback (RLHF) frameworks.
Industry Dive
Amazon Web Services (AWS)/Perplexity AI
AWS has (reportedly) launched an investigation into web scraping accusations involving Perplexity AI. Complaints have emerged that the AI-powered search engine gained unauthorized access to websites to scrape data.
Additionally, AWS is probing potential violations by AI firms per data training.
Character AI
On June 27, conversational tech firm Character AI launched human-AI character voice call functionality. The feature allows users to chat with AI avatars by voice and text. Additionally, users can create their (own) voice within its ecosystem.
Baidu
Chinese search engine giant Baidu introduced Ernie 4.0 Turbo, the latest model of its LLM. The move follows the October 2023 launch of the solution.
Furthermore, Baidu Chief Technology Officer Wang Haifeng iterated at an event that the model hit a 300 million-strong user base with support for 14.65 million developers and 370,000 corporate customers.
Everything Else
University of Reading (UK)
The rise of AI tech is a challenge to examinations. At least, that's what a study from the University of Reading (UK) suggests. Researchers fabricated 33 artificial students and used ChatGPT to answer undergraduate psychology exam questions.
The fake students scored (on average) half a point than their human counterparts. Additionally, the artificial answers were (relatively) undetectable at 94%, with researchers identifying the 9% gap as higher than usual (overestimated).
Bill Gates
At the Breakthrough Energy Summit in London, tech pioneer Bill Gates allayed concerns about increasing energy usage by AI ecosystems. Gates indicated that AI technology acceleration will lead to a reduction in energy emissions.
Consequently, Gates (also) identified the (increased) readiness of tech companies to pay for cleaner energy sources and called for pure intentions in developing AI technologies.
Open AI/Microsoft/Google
According to a June 28 NBC News report, several AI chatbots spread false (debunked) rumors about the June 27 presidential election debate. Leading chatbots, including Open AI's ChatGPT, Microsoft's Gemini, and Google's Gemini, confirmed inaccurate details about a (rumored) delay of the said event.
Following the Money
Hebbia
Knowledge-focused AI startup Hebbia (successfully) raised $100 million in a Series B funding round. Andreessen Horowitz led the round, valuing the firm between $700 million and $800 million.
Nokia/Infinera
To expand its American presence, Nokia concluded plans to acquire network infrastructure firm Infinera in a $2.3 billion deal. The acquisition is cash-and-stock-based and values Infinera at a 37% premium.
Etched
According to a June 25 CNBC report, two-year-old AI hardware startup Etched raised $120 million in a Series A funding round. The Cupertino, California-based startup is working on Sohu, an AI chip that focuses on transformers, an emerging AI infrastructure tech that is efficient in data training and deployment.
RECEIVE THE DAILY AI 3 TIMES PER WEEK TO YOUR INBOX. SUBSCRIBE NOW
ABOUT The Daily AI™:
The Daily AI™ is a new paid newsletter published every M/W/F from the team at Digital Wealth News™ and AI & Finance™ covering the artificial intelligence ecosystem worldwide. Subscribe today to stay abreast of all the latest news and players in the AI/ML sector.
**Disclaimer**: The information on the Daily AI™ is for general informational and educational purposes only. It is not intended as investment advice and should not be relied upon as a substitute for consultations with qualified professionals familiar with your financial needs. The Daily AI™ does not endorse specific investments, strategies, or financial service firms. The opinions expressed by authors do not necessarily reflect the views of the Daily AI™ or its affiliated brands or entities. Please conduct your (own) research and due diligence, and consult your financial advisor before making investment decisions.