- Future Perfect
- ✨86. 🤖14 LLMs besides ChatGPT | 🥵hottest day globally in history | 🔋Toyota's 745 mi 10 min charging battery
✨86. 🤖14 LLMs besides ChatGPT | 🥵hottest day globally in history | 🔋Toyota's 745 mi 10 min charging battery
PLUS: AI tackles ancient languages ✨ AWS exec downplays AI risk ✨ H2-powered ferry to sail SF Bay ✨ UAE invests $54B in clean energy ✨ Tesla wants to pay you to drive ✨ China's WeRide gets self-driving license in UAE
🤖 ARTIFICIAL INTELLIGENCE
The AI industry is witnessing a surge in large language models (LLMs) that compete with OpenAI's flagship ChatGPT, each with their unique capabilities, strengths, and weaknesses. Some could even be better suited for specific applications.
Measuring LLMs' performance is complex due to the vast scope of these models and their usage diversity. Comparing them using resources like Hugging Face's Open LLM Leaderboard could be insightful but not entirely accurate. Meanwhile, platforms like OpenLLM and FastChat facilitate easy switching between different models.
Building an LLM is costly and time-consuming, involving data collection, processing through expensive hardware, and then producing the model. Monetizing and sustaining this work remains an open question. Some organizations are exploring open sourcing their results, while others rely on service-based billing models.
The article highlights 14 LLMs: Llama, Alpaca, Vicuna, NodePad, Orca, Jasper, Claude, Cerebras, Falcon, ImageBind, Gorilla, Ora.ai, AgentGPT, and FrugalGPT. Each LLM is unique, catering to different user needs and application areas, from producing content and facilitating chat to generating diverse data types and enabling cost efficiency.
Ultimately, the best LLM depends on specific project requirements. Testing these models with prompts and carefully evaluating their results will help users determine the right fit for their needs.
Researchers are utilizing artificial intelligence (AI) to translate ancient texts and languages, such as Cuneiform and Egyptian hieroglyphs, into English.
The main challenge for this endeavor is the scarcity of data, but researchers managed to gather tens of thousands of examples to train the AI models.
AI developers detail in an Oxford Academic report how they used natural language processing (NLP) to translate cuneiform tablets inscribed in the Akkadian language, a common language in ancient Middle East and Mesopotamia.
Another project, Google Fabricius, employs online tools to decode ancient Egyptian hieroglyphs into English.
The AI model called Akkademia, used to train the Babylonian Engine AI for translation, is open-source and accessible on the project's Github.
The Babylonian Engine aims to conduct translations for various ancient languages, starting with Akkadian, making ancient history more accessible.
Matt Wood, VP of Product at AWS, dismisses fears of AI posing an existential threat to humanity, describing AI as a "mathematical parlor trick" capable of generating and synthesizing information to help humans make better decisions.
AWS has been actively involved in AI, with its Sagemaker suite and, recently, Amazon Bedrock, a set of generative AI tools that help organizations build, train, fine-tune, and deploy large language models (LLMs).
Generative AI, particularly transformer models, can process complicated natural language inputs and map them to complex outputs, which has sparked customer engagement and excitement.
Large language models are used in enterprise use cases for a variety of tasks such as text generation, summation, image creation, semantic scoring, ranking, and personalization.
Amazon Bedrock service provides multiple LLM options from different vendors and can be used alongside Langchain to chain and sequence prompts across different models, allowing organizations to create complex systems.
Wood emphasizes the importance of understanding both the technical and cultural aspects when adopting new technologies like generative AI to truly innovate.
⚡️ POWER PROMPTING!
Power Prompting is an awesome new section of Future Perfect showcasing powerful prompt engineering techniques, hacks, and strategies! Here’s today’s Power Prompt!
I'm the CEO of a medium-sized company that specializes in [describe your company's specialty]. Recently, we've been facing challenges due to [describe challenges]. Our current strategy involves [describe current strategy], but it doesn't seem to be yielding the results we'd hoped for. We're also considering potential future trends such as [describe future trends]. Could you help me conduct a SWOT (Strengths, Weaknesses, Opportunities, Threats) analysis for our company based on the information provided and recommend a detailed strategic plan that addresses our current challenges and aligns with potential future trends in the industry?
Want to submit your own? Tweet me @MarcHoag!
If you DON'T like Power Prompts, why don't you?
I want to understand how I can improve this.
🤖🔥🤯 COOL AI TOOLS, APPS, VIDEOS, PODCASTS, LINKS, AND MORE!
🌡️ CLIMATE CHANGE & CLEAN ENERGY
Monday marked the world's hottest day on record, with an average temperature exceeding 17 degrees Celsius (62.6 degrees Fahrenheit) for the first time, according to the US National Oceanic and Atmospheric Administration (NOAA).
This record surpasses the previous daily high of 16.92 degrees Celsius, set on July 24 last year.
While the record has yet to be confirmed by other measurements, it is likely to be surpassed soon as the northern hemisphere's summer continues, with global temperatures typically rising until the end of July or early August.
Last month, average global temperatures were the warmest ever recorded for the start of June, according to the European Union's Copernicus climate monitoring unit.
The World Meteorological Organization confirmed on Monday that an El Nino weather phenomenon is now underway in the Pacific Ocean, likely to push temperatures further above historical averages over the next year.
The burning of fossil fuels continues to emit approximately 40 billion tonnes of CO2 into the atmosphere annually, contributing to global warming.
The San Francisco Bay Area Water Emergency Transportation Authority (WETA) is set to launch a hydrogen fuel cell-powered ferry, the Sea Change, in August 2023 as part of a six-month pilot project aimed at exploring the viability of hydrogen technology for passenger ferries.
The $14 million ferry will operate between Pier 41 and the Downtown San Francisco Ferry Terminal, making four round trips per day, with a ticket price of $1 each way.
The Sea Change, which generates only water vapor and condensed water as emissions, will be the first hydrogen-powered commercial maritime vessel in the United States.
The vessel, owned by Switch Maritime and leased to WETA for the pilot project's duration, is part of the larger Sea Change Hydrogen Ferry Demonstration Project, a public-private partnership.
The initiative is backed by several entities, including the NBA Champion Golden State Warriors and the Golden Gate Bridge, Highway and Transportation District.
While there are challenges surrounding limited access to hydrogen fuel, the Sea Change represents a unique opportunity for greener maritime transportation, with the potential to revolutionize the industry worldwide.
The United Arab Emirates (UAE) plans to triple its renewable energy supply and invest $40-$54 billion over the next seven years to meet its growing energy demands.
The strategy, announced by Sheikh Mohammed bin Rashid Al Maktoum, includes investments in low-emission hydrogen fuel and the development of infrastructure for electric vehicles.
The UAE aims to be carbon-neutral by 2050 and will host the COP28 climate summit later this year.
The announcement included the creation of an Investment Ministry, to be led by Mohamed Hassan Alsuwaidi, deputy chairman of clean energy firm Masdar.
Sultan al-Jaber, chairman of Masdar and CEO of the Abu Dhabi National Oil Co, will serve as president of the upcoming climate summit, drawing criticism from environmentalists.
Despite these efforts, the UAE remains heavily reliant on natural gas, with renewable sources currently accounting for a small portion of its energy supply.
🚗 AUTONOMOUS & ELECTRIC VEHICLES
Tesla is launching a summer hiring spree for drivers to gather data to improve vehicle performance.
The automaker is seeking "Seasonal Vehicle Operators" across North America for a three-month testing period.
The temporary roles don't require any specific educational qualifications but do demand a clean driving record, safe driving habits, and a minimum of four years of licensed driving experience.
The positions are available in multiple locations across the US, including Illinois, Minnesota, Utah, Texas, Colorado, Washington, California, Florida, Georgia, New York, Arizona, and Massachusetts.
There are also several job postings in Canada, in Ontario and Quebec.
The pay for these positions ranges between $20 and $29 per hour, according to Glassdoor.
Toyota claims it has made a breakthrough in solid-state battery technology that could cut weight, size, and cost of batteries by half.
The company states it has developed a solid-state battery with a range of 745 miles that can charge in less than 10 minutes and is simpler to manufacture than traditional lithium-ion batteries.
Skepticism remains high due to the lack of specific details regarding energy density, charge and discharge rates, the number of charging cycles possible, performance in cold temperatures, and material composition.
Toyota is still committing significant resources to hydrogen fuel cell technology, despite this new development.
The company is cautious in its BEV production projections, planning to manufacture 3 million units annually by 2030, half of them with solid-state batteries.
The United Arab Emirates (UAE) has granted Chinese autonomous vehicle company WeRide the country's first national license for self-driving vehicles, permitting the company to test its Level 4 autonomous vehicles on public roads.
This move aligns with UAE ruler Sheikh Mohammed bin Rashid Al Maktoum's vision to have 25% of the country's transportation fully autonomous by 2030.
WeRide's permit was approved alongside a national policy for electric vehicles, which involves building a national charging network and promoting related industries to reduce emissions.
WeRide plans to test various types of self-driving vehicles and aims to commercialize its technology across a range of vehicles, including robotaxis, robobuses, robovans, and autonomous street sweepers.
WeRide has already been testing robotaxis on certain public roads in the UAE and has set its sights on expansion in Saudi Arabia.
The regulations for testing, deploying, and commercializing autonomous vehicles in the UAE are currently unclear, with the testing process to be carried out by the RegLab, an initiative by the General Secretariat of the Cabinet.
🎉 THAT’S ALL FOR TODAY!
What'd you think of today's email?
This helps me improve it for you!
✅ Share this post if you liked it
✅ Share this link (https://www.futureperfect.news) if you're a fan of Future Perfect
✅ Vote above before leaving!
Thanks for reading!
Looking for past newsletters? You can find them all here.