#47. ⚠️ AI is more dangerous than climate change | 🔋 Rimac Energy's grid-scale storage | 🚗 Tesla's end-to-end AI in Autopilot v12

✨ NEW! LINKS TO COOL AI STUFF!

Now that Future Perfect includes a whole section on AI, I realized it made sense to include a collection of three awesome AI-related links in every email, immediately after the AI section below. Enjoy!

🤖 ARTIFICIAL INTELLIGENCE

ChatGPT Is Powered by Human Contractors Getting Paid $15 Per Hour (short read)

ChatGPT, the renowned AI chatbot, relies significantly on the efforts of low-paid human workers who help guide its automation through data labelling. The OpenAI company, responsible for ChatGPT, compensates these U.S. contractors with $15 per hour for their essential role in shaping the bot's responses. Despite the importance of their task in training machine learning models, these workers, often referred to as moderators, receive no benefits and are paid nearly the minimum wage in some states. OpenAI was previously criticized for outsourcing this work to African moderators, paying them as low as $2 per hour. This reliance on underpaid human labor challenges the perception of AI as a self-sustaining system, underlining the need for better recognition and compensation for these "invisible" contributors.

AI pioneer says threat may overshadow climate change (short read)

Geoffrey Hinton, a distinguished pioneer in artificial intelligence, has expressed concerns that AI could potentially pose a more immediate existential threat than climate change. While acknowledging the gravity of climate change, he suggests that solutions to AI risks are less clear-cut. Hinton, a former Google executive, advocates for serious consideration of measures to prevent AI from gaining control over humanity, stressing the danger when "smart things can outsmart us." This perspective is shared by other computer scientists, including Microsoft's chief economist, Michael Schwarz, who warns that AI could be exploited by malicious actors to cause significant harm.

Here are dozens of ways AI could be used for harm — and some too scary to test (medium read)

A London-based team has delved into the potential misuse of large language models (LLMs) like GPT-4, suggesting a range of harmful applications from scientific fraud to manipulative AI bots. Their findings, collected under the website "LLMs are Going Great! (LAGG)", highlight "grey-area" use cases that could be challenging to regulate. Examples include a chatbot designed to manipulate users' opinions subtly and the creation of fake identities or credentials. They also noted the ease with which AI can generate misleading information, like bogus academic papers or fake references. The group also explored more whimsical, though potentially intrusive, ideas such as an Amazon Alexa programmed to play AI-generated Taylor Swift songs triggered by a secret word.

🤖🔗🤯 COOL AI TOOLS, APPS, VIDEOS, PODCASTS, LINKS, AND MORE!

• The A.I. Dilemma [VIDEO]: Center for Humane Technology Co-Founders talk about the profound risks of AI like ChatGPT.

• Inflection AI's Pi: A "personal intelligence" ChatGPT-like chatbot; I think this thing is incredible, especially it's text-to-speech capabilities which are basically completely human-sounding.

• Perplexity: A conversational search engine via a chatbot that answers with citations.

🌡️ CLIMATE CHANGE & CLEAN ENERGY

The US offshore wind’s pipeline is now 51+ GW – here’s where it’s happening (short read)

The U.S. offshore wind industry is expanding rapidly, with a pipeline comprising over 51,377 megawatts (MW) of anticipated capacity across 32 leases, enough to power more than 20 million homes. According to the American Clean Power Association, this growth is expected to create up to 83,000 jobs by 2030. The East Coast is the central hub, accounting for 84% of the pipeline. Investment in the offshore wind industry exceeds $1.7 billion, and the domestic supply chain is projected to grow significantly. However, rising project costs due to inflation, supply chain issues, and other economic disruptions pose challenges to the industry's economic viability, with steel costs being a primary concern. Collaborative efforts between industry stakeholders and government bodies are needed to overcome these hurdles and fully realize offshore wind's potential in the clean energy future.

Rimac Energy Plans To Add Grid-Scale Storage To Its Portfolio This Year (medium read)

Rimac, known for its expertise in electric vehicle technology, is planning to extend its portfolio by adding grid-scale energy storage this year. The company is launching Rimac Energy, aiming to accelerate the transition towards sustainable power and leverage its market-leading electric vehicle technology to create next-generation stationary Energy Storage Systems (ESS). The systems, which will be fully designed, developed, and produced in Europe, will reportedly feature a unique electrical architecture enabling significant performance, safety, and availability advantages. Pilot systems will be created for selected customers this year, with commissioning set for 2024 and high volume ESS production starting in 2025 at the Rimac Campus in Croatia. The company plans to scale up to more than 10 gigawatt-hours (GWh) of annual production.

Climate change: Vietnam records highest-ever temperature of 44.1C (short read)

Vietnam has recorded its highest ever temperature of 44.1C (111F), and experts predict that this record will soon be surpassed due to climate change. The record was set in the northern province of Thanh Hoa, with officials warning residents to stay indoors during the peak heat. This extreme heat has been experienced across the region, with Thailand reporting a record-equalling 44.6C in its western Mak province, and Myanmar's media reporting a decade-high temperature of 43.8C in the east. Climate change experts warn that such records will continue to be broken as the world has already warmed by about 1.1C since the industrial era began and temperatures will continue to rise unless governments enact significant emissions reductions.

🚗 AUTONOMOUS & ELECTRIC VEHICLES

Fisker (FSR) delivers the first Ocean electric SUV model (short read)

American electric vehicle manufacturer Fisker Inc. has delivered its first Ocean SUV model to a customer in Copenhagen, Denmark. The Ocean is the first electric vehicle from Fisker since the company's restructuring in 2016. The company has seen significant interest in the Ocean SUV since its public debut at the 2022 LA Auto Show, with over 63,000 orders reported in November. The vehicle, which has been certified by European regulators, is expected to offer the highest driving range of any battery-electric SUV currently available in Europe, with up to 707 km (439 miles) per charge. Fisker plans to deliver all Ocean One models by September 2023, with deliveries of the Ocean Extreme also set to begin in the same month. The company has not yet announced when US deliveries will begin.

Deploying Electric Buses At Airports Is Easy With This Tool (long read)

The US National Renewable Energy Laboratory (NREL) has developed a simulation-based optimization model to design effective electric bus deployment strategies at airports. The model was used at Dallas Fort Worth International Airport (DFW) to analyze real-world conditions and simulate the existing DFW system to determine optimal battery capacity, charging power, and number of charging stations while minimizing capital costs and emissions. The simulations represented many configurations, including battery capacity, charging power, placement and number of charging stations, travel distance of non-electric buses, passenger wait times, and other key factors. The results showed how these configurations affected cost, non-electric bus travel distance, and emissions. The model provides options based on combinations of factors and allows shuttle operations managers to consider those factors in planning. The research is part of a larger effort by NREL to decarbonize airport shuttle operations.

Tesla changes course with Full Self-Driving to introduce end-to-end AI in v12, says Musk (short read)

Tesla is planning to introduce end-to-end artificial intelligence (AI) in its Full Self-Driving (FSD) Beta version 12, according to CEO Elon Musk. The current FSD Beta uses AI and neural networks for perception, helping the vehicle detect its environment. However, for decision-making and driving input, the company relies on traditional programming. Musk had previously stated that AI wouldn't be necessary for these components, but his stance has since shifted. The upcoming v12 update, which doesn't have a specified release date yet, will integrate end-to-end AI for improved driving functionality. Musk has stated that he believes Tesla can achieve "full autonomy" this year, although it's unclear what he means by this statement.

🎉 THAT’S ALL FOR TODAY! PLEASE VOTE BELOW BEFORE LEAVING!

Thanks for reading! If you're a fan of Future Perfect, please be sure to share it with your friends and colleagues with this link.

Thanks! 🙏

-Marc 👋

Looking for past newsletters? You can find them all here.