![Real uses for AI in crypto, Google’s GPT-4 rival, AI edge for bad employees – Cointelegraph Magazine Real uses for AI in crypto, Google’s GPT-4 rival, AI edge for bad employees – Cointelegraph Magazine](https://i0.wp.com/cointelegraph.com/magazine/wp-content/uploads/2023/08/magazine-AI-Eye-3008-scaled.jpg?fit=%2C&ssl=1)
Artificial Intelligence and Cryptocurrency Are More Than Just a Buzzword
AI Eye has been making appearances at South Korea Blockchain Week and Singapore Token2049 over the past two weeks, trying to understand how cryptocurrency project leaders plan to use artificial intelligence.
The most famous is probably Maker founder Rune Christensen, who essentially plans to relaunch his decade-old project as a bunch of child DAOs governed by artificial intelligence.
“People misunderstand what AI governance means, right? We’re not talking about AI running a DAO,” he said, adding that AI won’t enforce any rules. “AI can’t do that because it’s unreliable.” Instead, the project focuses on using AI for coordination and communication—as an “atlas,” as they say, for the entire project.
“Having this central data repository makes it possible for hundreds of thousands of people from different backgrounds and different levels of understanding to meaningfully collaborate and interact because they have this shared language.”
Recent founder Illia Polosukhin may be better known in artificial intelligence circles because his project started as an artificial intelligence startup and later moved to blockchain. Polosukhin is one of the authors of the 2017 Transformer paper (“Attention Is All You Need”), which laid the foundation for the explosion of generative AI such as ChatGPT last year.
Polosukhin has too many ideas for legitimate AI use cases in the crypto space to detail here, but one idea he’s very passionate about is using blockchain to prove the origin of content so users can distinguish real content from AI-generated nonsense. Such a system would use cryptography to include provenance and reputation.
![Ilya Polosukhin](https://i0.wp.com/cointelegraph.com/magazine/wp-content/uploads/2023/09/photo_2023-09-19_14-01-31-1024x768.jpg?resize=640%2C481&ssl=1)
“So cryptography is like a tool to ensure consistency and traceability. And then you need the reputation around that cryptography, the on-chain accounts and record keeping, to really ensure that (X) published this content and (X) is now working for Cointelegraph.”
Sebastien Borget of The Sandbox said the platform has been using artificial intelligence for content moderation over the past year. “In-game dialogue in any language is actually filtered so it’s no longer toxic,” he explained. The project is also looking into its use for music and avatar generation, as well as more general user-generated content for world-building.
Meanwhile, Framework Ventures founder Vance Spencer outlined four main use cases for artificial intelligence, the most interesting of which is training artificial intelligence models and then selling them as on-chain tokens. Fortunately, Frameworks has invested in a game called AI Arena, where players can train AI models to participate in the game.
Keep an eye out for the magazine’s in-depth feature, which outlines their ideas in more detail.
Is artificial intelligence at the service of communists?
When it comes to artificial intelligence and cryptocurrencies, are they going in opposite directions? Dynamo Dao’s Patrick Scott digs into PayPal founder Peter Thiel’s thoughts on artificial intelligence and cryptocurrency in the foreword to the 1997 non-fiction book reissue sovereign individual, Cryptocurrencies and more are predicted in it. In it, Thiel argued that artificial intelligence is a technology of control, while cryptocurrency is a technology of liberation.
“In theory, artificial intelligence could make it possible to centralize control of the entire economy. It is no coincidence that artificial intelligence has become the Chinese Communist Party’s favorite technology. On the other hand, strong cryptography brings the prospect of a decentralized and personalized world . If artificial intelligence is communist, then cryptocurrencies are libertarian.”
Excerpted from Peter Thiel’s preface to the 2020 new work “The Sovereign Individual”. pic.twitter.com/cWzfgyUqfY
— Patrick Scott | Dynamo DeFi (@Dynamo_Patrick) September 17, 2023
Roblox lets users build with AI
Roblox has launched a new feature called Assistant that will allow users to create and code virtual assets using generative AI. In the demo, users write things like “make a game set in ancient ruins” and “add some trees” and the artificial intelligence does the rest. It’s still in development and will be released late this year or early next year. The plan is for Assistant to one day be able to generate complex gameplay or create 3D models from scratch.
![Roblox](https://i0.wp.com/cointelegraph.com/magazine/wp-content/uploads/2023/09/roblox.jpg?resize=525%2C458&ssl=1)
Bad employees benefit most from artificial intelligence
The worst employees in your workplace may benefit the most from using artificial intelligence tools, according to A new study from Boston Consulting Group. After using artificial intelligence, the output of below-average workers increased by 43%, while the output of above-average workers increased by only 17%.
Interestingly, employees who use AI to handle things beyond their current capabilities perform 20% worse because the AI feeds them responses that appear reasonable but are incorrect.
Google Gemini ready for release
Google’s GPT-4 competitor is about to be released, and according to The Information, a small group of companies have received early access to Gemini. For those who joined in late, Google was seen as leading the AI race until last November when OpenAI brought ChatGPT to market (arguably before it was ready) and catapulted it into the lead.
Google wants Geminis to be the best they can be GPT-4 Not only does it offer text generation, it also offers image generation, enabling the creation of contextual images (rumor has it that it is being trained on material such as YouTube content). There are plans to introduce features in the future, such as using it to control software or analyze charts via voice. Google co-founder Sergey Brin, who is said to have played an important role in the evaluation and training of the model, emphasized the importance of Gemini.
Also read
feature
Should we ban ransomware payments?It’s an attractive but dangerous idea
feature
Escape from LA: Why Sri Lanka’s lockdown worked for MyEtherWallet founder
AI expert Brian Roemmele said he has been testing a version of Gemini and found that it is “equivalent to ChatGPT-4, but with the addition of a second knowledge base. This saves it from some of the hallucinations.”
Google CEO Sundar Pichai told Wired this week that he didn’t regret not launching a chatbot sooner to defeat ChatGPT because the technology “wasn’t available until we implemented it into our products.” Needed to be more mature before that.”
“I’m not entirely sure if that would work,” Pichai said. “The truth is, after people see how it works, we can do a lot more. It really won’t matter in the next five to 10 years.”
AI meets 15-minute city
Researchers A professor at China’s Tsinghua University has built an artificial intelligence system that can plan cities based on current ideas about walkable “15-minute cities” with lots of green space (please send your conspiracy theories on this topic directly to X).
Researchers have found that artificial intelligence is better at tedious calculations and repetitive tasks, able to complete in seconds what would take a human planner 50 to 100 minutes. Overall, they determined that it improved user-friendly design by 50% when assessing levels of services, green space and traffic.
The headline numbers are a bit misleading, though, as the final plan only increases access to basic services by 12% and parks by 5%. In a blind review process, 100 city planners showed a clear preference for certain AI designs but no preference for others. The researchers envision their artificial intelligence as an assistant doing boring things while humans focus on more challenging and creative aspects.
Stephen Fry was cloned
blackadder and quality index The star of beloved British comedy institution Stephen Fry says he has fallen victim to artificial intelligence voice cloning.
On September 14, Frye played clip Apparently, he narrated a historical documentary at London’s CogX festival last week, but revealed the voice wasn’t his at all. “I didn’t say a word about it – it’s a machine,” he said. “They took the seven Harry Potter books that I read and created an AI of my voice based on that data set and created a new narration.”
Training artificial intelligence to steal actors’ work and repurpose it elsewhere without payment is one of the key issues in the current actor and writer’s strike in Hollywood. Frye said that this incident is just the tip of the iceberg and that artificial intelligence will “advance faster than any technology we have ever seen. We can all agree on one thing: It’s really weird to be alive right now.”
![quality index](https://i0.wp.com/cointelegraph.com/magazine/wp-content/uploads/2023/09/Stephen-Fry.webp?resize=640%2C360&ssl=1)
How to use ChatGPT to avoid cheating
The academics who enjoy using ChatGPT to cheat appear to be the ones who make stupid mistakes and leak this fact.Papers published in journals written physics The text was retracted after computer scientist Guillaume Cabanac noticed “re-responses” in the text, indicating that it was copied directly from ChatGPT.
Since 2015, Cabanac has helped discover hundreds of AI-generated academic manuscripts, including a paper in the August edition resource policywhich contains the sentence: “Please note that as an artificial intelligence language model, I cannot…”
![written physics](https://i0.wp.com/cointelegraph.com/magazine/wp-content/uploads/2023/09/Paper.jpg?resize=557%2C472&ssl=1)
All Killer No Filler Artificial Intelligence News
According to the Wall Street Journal, Meta is also developing a new model to compete with GPT-4, which is planned to be launched in 2024. It is designed to be many times more powerful than the existing Llama 2.
— Microsoft open sourced a new protein-generating artificial intelligence called Evolution differential. It works similarly to stabilized diffusion and Dall-E2, but instead of generating images, it designs proteins that can be used for specific medical purposes. This is expected to lead to new classes of drugs and therapies.
— Defense contractor Palantir has signed on, along with Cohere, IBM, Nvidia, Salesforce, Scale AI and Stability white housePlans for responsible AI development are somewhat vague. The administration is also working on an executive order on artificial intelligence and plans to introduce bipartisan legislation.
— Recently, 60 U.S. senators attended a private briefing on the risks of artificial intelligence with 20 Silicon Valley execs and nerds, including Sam Altman, Mark Zuckerberg Zuckerberg and Bill Gates. Elon Musk later told reporters that the meeting “will probably go down in history as being very important to the future of civilization.”
— ChatGPT traffic has Fallen It has declined for three consecutive months, with a decline of about 10% in June and July, and a further decline of 3.2% in August. The average time users spend on the site dropped from an average of 8.7 minutes in March to 7 minutes last month.
— Finnish prisoners are getting paid $1.67 Helping a startup called Metroc train artificial intelligence models. Artificial intelligence is learning how to determine when construction projects are hiring.
— The United States is far ahead in the artificial intelligence race, with 4,643 new startups and $249 billion in investment since 2013, 1.9 times the number of startups in China and Europe combined.
Also read
feature
When Worlds Collide: Joining Web3 and Cryptocurrencies from Web2
List
Wall Street disaster expert Bill Noble: Cryptocurrency spring is inevitable
Video of the week
Writer and storyteller Jon Finger tried the HeyGen video app, which not only translated his words but also cloned his voice and synchronized his lip movements with the translated text.
Test it out @HeyGen_Official French and German translation. I don’t speak either language, so if you do, please let me know if this sounds natural.
I wish you could turn off color correction if you paid.
It doesn’t work on my phone so I have to upload to my computer. pic.twitter.com/iF5eONAQ3c— Jon Finger (@mrjonfinger) September 11, 2023
subscription
The most fascinating read in blockchain. Delivery is weekly.
![Subscribe to the Cointelegraph newsletter.](https://i0.wp.com/cointelegraph.com/magazine/wp-content/uploads/2022/10/reading-copy.png?w=640&ssl=1)
Svlook