- The AI Pulse
- Posts
- š¤ Amazon Invests Over $500 Million Into Nuclear Power
š¤ Amazon Invests Over $500 Million Into Nuclear Power
PLUS: The New York Times Takes Legal Action Against Perplexity AI, Metaās Scientists Develop āThinkingā LLMs
Welcome back AI enthusiasts!
In todayās AI Report:
šµAmazon Invests Over $500 Million Into Nuclear Power
šThe New York Times Takes Legal Action Against Perplexity AI
šMeta Scientists Develop āThinkingā LLMs
š Trending Tools
š°Funding Frontlines
š¼Whoās Hiring?
Read Time: 3 minutes
šRECENT NEWS
AMAZON
šµAmazon Invests Over $500 Million Into Nuclear Power
Image Source: Canvaās AI Image Generators/Magic Media
Amazon Web Services (AWS), Amazonās cloud computing platform that deploys software solutions, is investing over $500 million into nuclear power.
Key Details:
AWS has signed an agreement with Dominion Energy, a utility company that offers clean, safe, reliable, and affordable energy solutions.
The agreement explores building Small Modular Reactors (SMRs) near Dominion Energyās existing North Anna Power Station.
SMRs are smaller, more efficient nuclear reactors designed to be built in factories and shipped to operational sites for installation.
SMRs boast faster construction times, allowing them to come online sooner.
AWS needs SMRs to power the growing demand for Generative AI (GenAI) platforms. āWeāre seeing the need for GigaWatts (GW) of power in the coming years,ā said AWS CEO Matt Garman.
GenAI uses AI models trained on text, image, audio, video, and code data to generate new content.
AWS plans to invest $35 billion by 2040 to establish multiple data center campuses across Virginia powered by SMRs.
Why Itās Important:
Big Tech companies are investing billions of dollars into nuclear power due to the growing demand for GenAI platforms.
Earlier this week, Google partnered with Kairos Power to build āseveral small nuclear-power reactorsā to fuel AI ambitions.
Last month, Microsoft signed a historic 20-year deal with Constellation Energy to reopen Three Mile Island near Middletown, PA, to access nuclear power for AI demands.
THE NEW YORK TIMES
š°The New York Times Takes Legal Action Against Perplexity AI
The New York Times just sent a ācease and desistā letter to Perplexity AI for violating copyright laws.
Key Details:
Perplexity AI is a free AI-powered search engine that rivals Google Search. In January, Jeff Bezos participated in a $73.6 million Series B funding round for Perplexity AI.
The New York Times claims Perplexity AI is leveraging the media outletās articles to develop AI-generated summaries without their permission, which violates copyright law.
āPerplexity AI and its business partners have been unjustly enriched by using, without authorization, the Timesās expressive, carefully written and researched, and edited journalism without a license,ā the media outlet wrote.
Perplexity AI says itās open to working with publishers and will respond to the ācease and desistā letter by the October 30th, 2024 deadline.
The New York Times also filed a lawsuit against Microsoft and OpenAI over abusing the newspaperās intellectual property to train Microsoftās Bing Search and OpenAIās ChatGPT.
Why Itās Important:
Microsoft, OpenAI, and Perplexity AI could potentially argue that AI models fall under transformative use (i.e., add new expression, meaning, or messaging to the original work) and bypass copyright law. The AI models learn patterns from the media outletās content rather than directly copying it.
They could cite Google LLC v. Oracle America, Inc., where the Supreme Court ruled that Googleās use of Oracleās Java API Code to create the Android platform was transformative use because it served a new purpose for a substantially different product, even though it involved copyrighted content.
š©ŗ PULSE CHECK
Do you think Perplexity AIās situation falls under transformative use?Vote Below to View Live Results |
AI RESEARCH
šMeta Scientists Develop āThinkingā LLMs
Image Source: Meta Fundamental AI Research (FAIR)/āThinking LLMs: General Instruction Following With Thought Generationā/Screenshot
Meta scientists developed Thought Preference Optimization (TPO), a method for training Large Language Models (LLMs) to āthinkā before answering questions.
LLMs are AI models pre-trained on vast amounts of data to generate human-like text, such as OpenAIās ChatGPT.
One of the main challenges with LLMs is their tendency to respond to questions without contemplating the complexity of the question. For simple questions, immediate responses may be sufficient. However, LLMs often fail to answer complex problems that require reasoning and problem-solving.
To solve this issue, Meta scientists hypothesized how to get LLMs to pause, create internal thoughts, and evaluate those internal thoughts before generating responses.
IPO achieves this by instructing LLMs to divide outputs into two parts:
Thought Process: Multiple thoughts are generated for each question and paired together.
Final Response: The Thought Process pairs are evaluated through preference optimization. The best Thought Process pairs are selected for further training iterations.
š TRENDING TOOLS
āļøCoinfeeds is your crypto chatbot companion.
šŖ¶Catbird is a visual brand generator for creatives.
šCrowdView is a search engine for forums and message boards.
šHypeIndex tracks stock market trends and investor sentiment.
šŗPlacerAI gathers consumer foot traffic insights on any property.
š®Browse our always Up-To-Date AI Tools Database.
š°FUNDING FRONTLINES
š¼WHOāS HIRING?
T-Mobile (Bellevue, WA): IT Software Engineering Intern, Summer 2025
XTX Markets (London, UK): Software Engineering Intern, Summer 2025
Adobe (San Jose, CA): Software Development Engineer, New College Grad 2025
Nvidia (Santa Clara, CA): Solution Architect, Worldwide Field Operations, New College Grad 2025
Koddi (Fort Worth, TX): Junior Software Engineering Cohort, Early Career
š¤PROMPT OF THE DAY
SALES SUPPORT
šŖCompelling Sales Pitch
Write a compelling sales pitch for [Small Business] with [Product/Service] in [Industry] tailored to [Target Customer].
Make sure itās well-structured, informative, and persuasive, leaving the [Target Customer] feeling confident and excited about the potential benefits of purchasing the [Product/Service].
Small Business = [Insert Here]
Product/Service = [Insert Here]
Industry = [Insert Here]
Target Customer = [Insert Here]
šFINAL NOTE
FEEDBACK
How would you rate todayās email?It helps us improve the content for you! |
ā¤ļøTAIP Review of The Day
āI canāt get over the fact youāre only 22 and producing such an amazing AI resource.ā
REFER & EARN
šYour Friends Learn, You Earn!
You currently have 0 referrals, only 1 away from receiving āļøUltimate Prompt Engineering Guide.
Refer 3 friends to learn how to š·āāļøBuild Custom Versions of OpenAIās ChatGPT.
Copy and paste this link to friends: https://theaipulse.beehiiv.com/subscribe?ref=PLACEHOLDER
Reply