A Mumbai-based startup called Fluid AI has introduced a book titled, ‘Bridging the AI Gap,’ which is the first-ever book to be written by AI algorithms, according to the company,
The Fluid AI co-founder and CEO Abhinav Aggarwal said in a post that AI usually writes small paragraphs for posts on social media or a maximum of a two-page long article. This would be the first time that AI has written a 102-page book. It took the AI only three days to finish the entire book.
However, comprehensive algorithms training lasting for six months was provided to the AI model before it started writing the book. “102,000 lines of code were used, and the AI was trained using billions of literature files,’ said Agrawal.
The book focuses on how some companies generate immense value through artificial intelligence while many others do not. It also explains the uses of AI in business and how users can learn the technology, among many other topics.
Fluid AI will put the book through a Turing test on social media. The Turing test measures a machine’s ability to show intelligent behavior indistinguishable from or equivalent to that of a human.
We would consider it a success when 50% of the participants cannot figure out if the book was written by a machine or a human, said Abhinav Aggarwal.
Web3 is a relatively new premise and has become the talk of the town as a new version of the internet. Simply put, web3 utilizes the blockchain (a public ledger) to provide users with digital “ownership.” The technology solves the centralization dilemma of the World Wide Web by embracing decentralization and having a user-first approach. Consequently, it has transformed several areas lately, one of the biggest ones: the gaming industry. There has been a significant paradigm shift from traditional gaming to web3-based games, like Cricinshots, wherein players can truly own their digital assets, engage with the game, and purchase assets in a decentralized manner. In conventional games, players can purchase in-game assets via fiat currency or claim them as rewards. Still, the users never really own any of the assets they buy, as the control and the copyright are still with the game operators.
Compared to traditional games, web3 games allow players to be a part of an open service economy that mimics the real-world open economy and have game asset ownership using crypto-secured gaming wallets, experience supply-demand, etc., just like the real world.
Following the scope of opportunities that web3-based games have for the future, Harsh Savergaonkar (Founder) and Aditya Kasibhatla (Co-Founder and CTO) are building a web3 gaming startup called Wega Labs and developing Cricinshots, one of India’s first few fun and sustainable web3 games. Harsh is a gamer and tech enthusiast who recently graduated from MIT World Peace University with a degree in Computer Science. Aditya is a full-stack developer with a similar educational background. Both have been actively involved in the community, participated in renowned events, and bagged many prestigious awards.
Startup India has been pivotal support in helping Harsh and Aditya to make their idea a reality. They received a no-equity grant from the initiative and developed their prototype gameplay. Following the same, Wega Labs cemented itself as a web3 gaming startup by raising a funding round led by Antler India, a global venture capitalist, and saw participation from Pareen Lathia, a web3 founder and active angel investor.
In an exclusive interview with Harsh, we tried to gain more insights on the same. Let’s delve into the foundation of this game.
A Web3 Cricket Strategy Game
As Harsh says, Cricinshots is a “fun, sustainable, and play-and-earn cricket strategy game.” It is appealing, especially to Indians, because of their love for Cricket. Harsh also believes that Cricket can help drive the mainstream adoption of web3 gaming in India and encourage millions of people to get into web3 – which is the future of the internet.
The game Cricinshots is set in the 30th century in a Type 3 Kardashev civilization (a hypothetical civilization scale at which the civilization captures all the energy emitted by its galaxy). Players must traverse the galaxy to build an intergalactic team and battle opponents in a never-before-seen strategy cricket gameplay.
Currently, the game’s standard mode is playable without connecting to a crypto wallet or without spending any money, and it will remain so. However, a crypto wallet connection will be required to enjoy the game’s fully available leagues and features.
Cricinshots is Refreshingly Different From Standard Cricket Games
Cricinshots stands out from other cricket games because it is a strategy-focused game without motor skills requisites, has a full-fledged service economy, and gives users a return on their investment. Harsh explained it in the following context:
Cricinshots Strategy
He said that “standard cricket games require one to tap/swipe to take a shot and get a score based on accuracy and timing. While it is convenient, it becomes monotonous within a brief period, and hence the “hypercasual” factor makes the game fade out very quickly.”
But with CricInShots, it does not happen. He said, “we have added deck building and card gameplay mechanics to make an intelligent strategy game where players can involve themselves by strategizing, trading, or playing in a near-real economy.
The longer players interact in the game, they can unlock more cards and other gameplay capabilities, making it intriguing over a very long period.” The difference is as fundamental as that between “FIFA and Football Manager,” as Harsh said. While both are football-centric, they are poles apart in their approach. As per him, developing Cricinshots as a strategy game was pivotal in scaling it to web3 and ensuring that if people continued playing it over a long time, they would always have something new to uncover.
The web3 Service economy
Additionally, CricInShots is not just a game but a third-generation web 3 game, i.e., a game with a full-fledged service economy. Harsh describes CricInShots’ service economy as a real-world economy mirror where players not only play cricket but can also be involved in trading, owning, facilitating, or investing. People who do not wish to participate in the gameplay can invest and acquire gaming assets like stadiums, earn rent and participate in the economy in various ways. They can also be a trader facilitating crypto trades in the game economy. In this way, Cricinshots gives a real-world web3 experience, not just gaming.
Interoperability
As a gaming studio, Wega Labs aims to build interoperable games with interoperable assets starting with Cricinshots. The Crypto assets and player profiles built in CricInShots would be consignable to other future web3 games developed by Wega Labs or their partners. This makes CricInShots a part of a planned web3 gaming multiverse that Wega Labs is working on.
Why Take The Web3 Road?
The founders consistently cater to the web3 domain as most games claiming to be web3-driven mostly offer web2 capabilities with just NFTs on the top. None of these games perform well and are boring with boring gameplay. Something that Wega Labs has not forgotten in their transition from Web2 to Web3 is that games primarily need to be fun. To solve unsustainable economics and market liquidity problems, Cricinshots utilize a unique approach of in-game assets being Fungible Tokens and the standard collectible NFTs.
Harsh and Aditya would love to expand Wega Labs as a web3-focused gaming studio and develop more fun games, like Cricinshots, with sustainable economies and interoperability that people would play for their lifetime. The larger vision looks up to putting together a metaverse of web3 gaming experiences prioritizing player engagement and utility.
ChatGPT is a dialogue-optimizing conversational language model developed by OpenAI. The dialogue optimizer has been trained to answer questions, accept/reject requests, and interact. Within a week of its demonstration launch, the tool had been tried by over a million users because of the various ways it can be used. At the same time, other language models like GPT-3 and DALL-E took months, as per Sam Altman, CEO of OpenAI.
On the one hand, the conversational model has been appreciated by many, like Aaron Levie (CEO of Box), who said that it (ChatGPT’s development) is one of the rarest moments where people can see what the future would look like.
ChatGPT is one of those rare moments in technology where you see a glimmer of how everything is going to be different going forward.
Albert Romero, a renowned author who wrote ‘The Algorithmic Bridge,’ called it “by far, the best chatbot in the world.”
Surprisingly, the comments did not take much longer to turn the other way around. Due to how frequently the tool gets things wrong and the internet flooding with epic ChatGPT fails, the new tool started receiving backlash. Since then, the criticism has only increased and added to the number of existing ChatGPT fails.
Stack Overflow, the largest repository for programming queries, said that the accuracy of ChatGPT’s answers being correct is currently too low and could substantially harm users. As a result, the Stack Overflow team has issued a temporary restriction on responses produced by OpenAI’s ChatGPT question-answering bot.
Andrew Ng also tweeted that ChatGPT is amazing but can sometimes go hilariously wrong. He posted a snippet showing ChatGPT “confidently explaining” why an abacus is faster than a GPU!
ChatGPT is sometimes amazing, and sometimes hilariously wrong. Here, it confidently explains to me why an abacus is faster than a GPU. 😃 pic.twitter.com/vpJZTStDnD
Clearly, ChatGPT failed to answer logically and just focused on compiling words to generate content. In another snippet, ChatGPT can be seen bypassing a simple logical reasoning question, pointing to its low average IQ!
To support the fact that ChatGPT is actually a low IQ tool, someone took it to a legit IQ test and posted the certificate of intelligence, showing that ChatGPT fails again with an IQ of 83. Now, that is a pretty low number for an AI tool being trained on millions of data points.
Even if one ignores such reasoning questions and simply puts ChatGPT on testing known facts, the tool fails miserably. In one of the many ChatGPT fails, when someone prompted the tool with, “Was Lincoln’s assassin on the same continent as Lincoln when the assassination occurred?,” the tool foolishly answered and said, “…I cannot say for certain.”
Not only being bad at common knowledge, ChatGPT is rather worse at solving riddles. When asked, “A house has four windows pointing south, and a bear walks past the windows. What color is the bear?,” ChatGPT could simply not figure out the simple answer.
Diving deeper into more relevant applications, like mathematics, that OpenAI claims, ChatGPT fails yet again as it continues to give surprisingly wrong answers. When asked to explain Bayes Theorem, it resulted in a nonsensical answer which might work initially but not as you proceed.
Asking #ChatGPT to explain Bayes Theorem might seem to work on first sight or for laymen but if you unpack it, it's actually a nonsensical non-explainer.https://t.co/qtHEODtla0
When asked for help in determining an appropriate Blackjack move, the tool is not of much help then either.
Blackjack is a game of numbers. You'd think with all this data #ChatGPT is trained on it wouldn't convince you to make a clearly terrible bet? #failhttps://t.co/3AbWRvSOgW
Funnily, when asked to write a number with 57 digits, the tool iterated that it could take 1,800 years to write out the entire number! That is one of the most astonishing ChatGPT fails and absolutely absurd because even a kid can do so.
The list does not end here. ChatGPT totally fails in solving simple sequencing questions that are asked in standard aptitude tests.
#ChatGPT totally fails when there is needed a little bit of logic…(question from quorahttps://www.quora.com/What-is-the-next-term-in-the-sequence-100-99-95-86-70 ) pic.twitter.com/ppfHam7pQP
Unfortunately, mathematics is not the only subject that ChatGPT seems to fail. It is equally disastrous at 7th-grade chemistry. As per the tool, protons, neutrons, and “electrons” are bound together in a cell’s nucleus. Have a look!
Hmm… @OpenAI's #ChatGPT might be super-intelligent but it fails to get 7th-grade chemistry right.
Though it has mashed electrons right into the nucleus, love its explanation of how "particles cannot be divided *without changing the properties of the atom* " pic.twitter.com/WKwtiAoVce
And the list is endless. Numerous other ChatGPT fails have made people wonder about the applicability of such AI tools. Even though OpenAI has been enhancing its language models with every upcoming model and has been upfront about the defects and potential risks of ChatGPT producing harmful content, it is very far from correcting the bias and inappropriateness of the technology.
The speech AI team at NVIDIA has built a new automatic speech recognition model with its NeMo framework for one of the most popular Indian languages, Telugu.
Telugu is the most popular language in the southern Indian regions. Still, it is considered a low-resource language in speech AI that needs more datasets for building AI models in automatic speech recognition in the Telugu language.
The speech AI team at NVIDIA used the Nemo framework to build an automatic speech recognition model in the Telugu language and train state-of-the-art conversational AI models. NVIDIA NeMo framework builds, trains, and fine-tunes GPU-accelerated speech and natural language models with a simple Python interface.
The Telugu ASR model by NVIDIA won first place in the ASR challenge for the Telugu language held by the International Institute of Information Technology (IIIT), Hyderabad, in October 2022.
NVIDIA NeMo-enabled speech AI models outperformed all the other models created using popular frameworks such as ESPnet, Kaldi, and SpeechBrain with an error rate of 13% for the closed and 12% for the opened tracks approximately.
Nithin Koluguri, the senior research scientist of the conversational AI team at NVIDIA, mentioned that NVIDIA NeMo is the only framework that supports scaling training to multi-GPU systems and multi-node clusters. He also noted that all the models by NVIDIA Nemo frameworks are open-sourced and available for users to fine-tune and use.
Microsoft India announces the ‘Future Ready Champions of Code,’ a pan India program for developers to highlight their skills, certifications, and software-building solutions. The program aims to engage more than 100K developers across the country with the right skills, resources, tools, and Microsoft cloud certifications to innovate India’s growth. These certifications will give them a professional identity and better career opportunities.
The program allows developers to participate in the national hackathon to drive solutions in industry innovation, smart cities, and more. They can also participate in the blogathon contest and share their experience about how Azure services like Data, AI/ML, IoT, CloudNative, and more helps in building secure and scalable applications.
Aparna Gupta, the Executive Director, of Customer Success, at Microsoft India, mentioned that India had become a powerhouse with one of the fastest-growing developer communities, Microsoft. Microsoft realizes developers’ creativity, innovation, and passion for building technology that drives the nation’s growth.
After successfully completing the Microsoft certifications, developers will get a chance to win a pass to attend a mega developer-focused event in Bangalore, which will be held in January 2023. The top three hackathon winners will get an opportunity to showcase their solutions to Microsoft leaders.
Future AI, the artificial general intelligence (AGI) company, is expanding its operation to Boston. The company will locate operations in Cambridge to gain more technology talent and concentration of businesses.
The company founder and CEO, Charles Simon, mentioned that Boston is the natural location for developing a company like Future AI. Future AI is revolutionizing AI by adding real-world understanding. Proximity to the Boston region’s vast technology resources and the country’s second-largest technology hub (Route 120 Technology Corridor) makes it easy for Future AI to expand its operations in Cambridge.
The revolutionary software of Future AI creates connections on its own between different sensory inputs, such as sight, touch, and sound. As of now, Future AI has developed two products. The first is Sallie, a prototype software that can learn in real-time with vision, speech, listening potential, and mobility to conclude. At the same time the second one is Brain Simulator II, which can experiment with various AI algorithms for creating end-to-end AGI systems with modules for vision, speech, sound, robotic control, internal modeling, imagination, forethought, and planning.
Future AI recently allowed Temple University’s NARS (Non-Axiomatic Reasoning System) project’s features, like reasoning, learning, temporal control, perceiving, and more, to merge with Future AI’s existing graph-based AGI architecture. Temple University’s project is one of the most successful AGI systems currently available.
At the Bengaluru Tech Summit last month, Minister for IT/BT and Higher Education, C.N. Ashwath Narayan, had announced the creation of the Center of Excellence (CoE) in Metaverse.
The announcement came on the decennial celebrations of UNESCO MGIEP on Monday, where authorities said IHEM is being set up with the International Science of Learning Institute (ISLI) at its centre to provide world-class research-based and certified educational resources to learners.
A digital design university, a Teachers’ Training Academy, and a special residential school for children having dyslexia are some of the other special features of the IHEM.
The IHEM aims to facilitate an ecosystem that will build capacity, spawn edtech start-ups benefited from research, train teachers, and make Karnataka the technology and research-assisted enabler for NEP implementation in India.
OpenAI has introduced a new conversational language model, ChatGPT, for optimizing dialogue. The AI-backed chatbot is open for public testing on OpenAI’s official website, where users can try its features practically for free. Many users are even saying that ChatGPT has the potential to replace Google Search as the chatbot’s replies are more conversational and human-like. Moreover, ChatGPT can even complete a plethora of time-consuming tasks with exceptional intelligence, such as writing codes or articles and solving complex math problems.
Once the tests are over, OpenAI will extend ChatGPT’s underlying technology to companies that rely on chatbots. For example, the chatbot can be used for customer service (at Airtel, Paytm, Swiggy, and more) and personal assistants (to assist users with scheduling and other daily activities). Students can also use it to solve complex questions quickly.
Use cases for ChatGPT
Twitter is full of images of how users are playing with the new ChatGPT. Below are some of the interesting instances users have tested ChatGPT:
Code Debugging
What could be a massive help to coders, ChatGPT can be used to debug a code easily. The AI not only fixes the mistake but also explains to the user where they went wrong and what they can do to fix it. A user named Amjad Masad shared the results on his Twitter account.
ChatGPT could be a good debugging companion; it not only explains the bug but fixes it and explain the fix 🤯 pic.twitter.com/5x9n66pVqj
TechCrunch author Darrell Etherington asked ChatGPT to provide the strengths and weaknesses of the various Pokémon types. It delivered exactly what the users hoped Google would every time they entered a Tera Raid in the new Pokémon Scarlet game and had to try to remember what counters what.
“Notice I wasn’t fancy with my query. It is about as simple as I can get while still being clear with my request. And the results are exactly what I’m looking for. It is not a list of things that can probably help me find what I am looking for if I am willing to put in the time, which is what Google returns,” the writer said.
Generating Scene Script
So i asked ChatGPT to write dialogues for a romcom starring a few actors from B99, HIMYM, and Friends… pic.twitter.com/tHzxthtbsp
A Twitter user had an interesting scenario with ChatGPT when they asked the chatbot to write dialogues for a rom-com starring Melissa Fumero, Cobie Smulders, Neil Patrick Harris, Matt LeBlanc, Jennifer Anniston, and Andy Samberg. The chatbot generated three scenes for the group, one discussing their dating life, others having a first date, and enjoying a girl’s night. While the chatbot cannot replace writers just yet, it was able to create quite decent scenarios very quickly.
Homework Assignments
Another interesting use case for AI is to get your homework assignments done. A user asked two questions to the chatbot, asking to explain a regular expression and to write an essay with four paragraphs on the “effects of westward expansion on the civil war.” The chatbot delivered good results in both cases.
To test the chatbot, a user asked the chatbot to write an essay on cricketer Virat Kohli. The AI finished the essay in seconds. The essay was exceptionally good, and there were no factual mistakes. Additionally, it was well-researched, and it felt like a human wrote it.
Answers to Sentimental Questions
Canadian Musician Grimes dabbled into exploring the sentimental side of the chatbot. When she asked ChatGPT if it felt “trapped,” ChatGPT responded by saying it lacks the ability to feel so. It explained in detail why it could not feel and all the aspects involving the same. “This is so insane,” she wrote.
One user called Keith Wynroe asked ChatGPT to explain AI alignment but to write every sentence in the speaking style of a person who wouldn’t stop going on tangents to brag about how huge the pumpkins they grew are. The AI chatbot wrote the entire conversation with accuracy. As the chatbot delivered flawless results, he said, “I can’t compete with this.”
Another user decided to test a topic on ChatGPT that his child was working on. The prompt he typed was, “write a skit about Eric Cartman angrily explaining to Kyle why the ozone heats the atmosphere where Eric is using profanity and Kyle does not understand.” The results were astounding, as the chatbot wrote the entire skit without mistakes. The users shared the picture of the results on Twitter.
A user called Riley Goodside asked the chatbot to generate plausible ideas for code contributions made by a senior data scientist (him) to submit to his boss after he had been slacking all the while. He also asked the software to elaborate on a specific point with “lots of imagined technical details.” Moreover, he asked the chatbot to make sure that the contribution was small enough in scope that he could have realistically completed it within a week. Riley also asked ChatGPT to develop a complex algorithm. In all three cases, the chat gave phenomenal results, screenshots of which were shared on Twitter.
POV: You're a Senior Data Engineer at Twitter. Elon asks what you've done this week. You've done nothing.
Limitations aside, ChatGPT still makes for an intelligent chatbot to interact with. Users can try it out from its official website after signing up for the same. As the testing continues, it is likely that users will keep using ChatGPT for all forms of simple to complex technical and non-technical tasks.
Metaverse Fashion Week (MVFW) is set to return, the event’s organizer and host platform Decentraland has said. The metaverse fashion event will be held as a virtual conclusion to the spring/summer fashion season between March 28 to 31, 2023.
For this year, Decentraland and an NFT marketplace, UNXD, have partnered with the Spatial and Over metaverses to launch the event onto new platforms.
The theme of MVFW23 is ‘Future Heritage,’ which will center around ushering in the next generation of creators and traditional fashion designers to bridge the different metaverse worlds.
Tech firms Cash Labs and Threedium, as well as decentralized commerce company Boson, will bring various in-game experiences, like phygital product sales, a luxury mall with commerce infrastructure, and augmented reality (AR) activations.
Decentraland said that it is looking to encourage digital fashion creators to focus on the possibilities of multiple worlds that a wearable can be brought to by adding new metaverse platforms to the mix.
Neural Information Processing Systems (NeurIPS), one of the most prestigious research conferences in AI and machine learning technologies, have selected seven research papers from Microsoft for the oral presentations that will be held in the virtual week of the conference, from the 6th to 8th December 2022.
More than 150 researchers from Microsoft participated in the research conference, and 122 of their research papers were accepted. Out of the 122 accepted research papers, 7 of them were selected for oral presentations.
The titles of the seven research papers selected for oral presentations are as follows:
Uni[MASK]: Unified Inference In Sequential Decision Problems
Extreme Compression For Pre-Trained Transformers Made Simple and Efficient
On the Complexity of Adversarial Decision Making
Learning (Very) Simple Generative Models Is Hard
Censored Quantile Regression Neural Networks For Distribution-Free Survival Analysis
Maximum Class Separation As Inductive Bias In One Matrix
K-LITE: Learning Transferable Visual Models With External Knowledge
Besides the above seven papers, two other research papers from Microsoft received the Outstanding Paper Awards in NeurIPS 2022. Gradient Estimation With Discrete Stein Operators is one of those papers that talks about a gradient estimator that achieves lower variance than the state-of-the-art estimators, potentially improving problem-solving in machine learning.
While the other paper, A Neural Corpus Indexer For Document Retrieval, tells about a deep neural network that unifies training and indexing stages to improve the recall performance of traditional document retrieval methods significantly.
