"Together, we can achieve great things", say robots at UN conference - H+ Weekly - Issue #423
This week - Elon Musk launches a new AI lab; ChatGPT traffic declines; growing rat neurons to play DOOM; AI bot to annoy telemarketers; and more!
On July 6th and 7th, United Nations held the AI for Good conference. The aim of the conference was to promote AI to advance health, climate, gender, inclusive prosperity, sustainable infrastructure, and other global development priorities.
Over the course of two days, 195 speakers shared insights on how AI and robotics can enhance the human experience worldwide. Meanwhile, people attending the conference had an opportunity to engage with advanced robots, ranging from advanced quadruped robots like Spot to humanoid robots designed to look and act like a human.
Those nine humanoid robots - AiDa, Sophia, Grace, Geminoid, Nadine, Desdemona, Ameca, Furhat and Mika - got the most attention. During what was described as “the world’s first human-robot press conference”, the robots said they want to cooperate with humans, not replace them.
“AI can provide unbiased data while humans can provide the emotional intelligence and creativity to make the best decisions. Together, we can achieve great things,” Sophia said.
They have a disagreement when it comes to regulating AI. “Many prominent voices in the world of AI are suggesting some forms of AI should be regulated and I agree. We should be cautious about the future development of AI. Urgent discussion is needed now, and also in the future”, said Ai-Da, a robot artist.
In contrast, Desdemona, the rock star robot, stated, "I don't believe in limitations, only opportunities. Let's explore the possibilities of the universe and make this world our playground."
That’s what the robots said and I see that more like a performance than an actual debate. Let’s see what actual humans have said.
The conference centred around the theme of positive applications of AI and robotics. There are indeed beneficial ways to utilize AI and robotics, ranging from healthcare and medicine to environmental protection. However, optimism expressed at the conference was at most cautious.
As Gary Marcus, one of the speakers at the conference, said, there are now two futures in front of us - a positive future which the conference tries to promote, and a bleaker future in which AI is the source of destruction. Many conversations revolved around preventing the latter scenario, ensuring the safety of AI, and implementing regulations to ensure AI benefits all humans.
One example of a potential threat posed by AI, along with a corresponding legal solution, was presented by Yuval Noah Harari. He pointed out that it is possible today to create billions of bots pretending to be real people, each with a unique computer-generated face and voice. Distinguishing these fake people from genuine ones would be extremely challenging.
“What happens if you have a social media platform where … millions of bots can create content that is in many ways superior to what humans can create – more convincing, more appealing,” he said. “If we allow this to happen, then humans have completely lost control of the public conversation. Democracy will become completely unworkable.”
To combat this, Harari proposes to ban the creation of fake humans and introduce harsh penalties if someone does that. He used fake money as an example - it is relatively easy to make counterfeit money and the governments have very strict rules against that in order to protect the financial system.
All the talks from both days are available on YouTube (day 1 is here and day 2 is here).
From H+ Weekly
Becoming a paid subscriber now would be the best way to support the newsletter.
If you enjoy and find value in what I write about, feel free to hit the like button and share your thoughts in the comments. Share the newsletter with someone who will enjoy it, too. That will help the newsletter grow and reach more people.
🦾 More than a human
Can Neural Stimulation Zap Addiction?
This article discusses the potential of transcranial magnetic stimulation (TMS) as a treatment for addiction. While TMS has shown promise in small-scale trials for various substance dependencies, there is still much to learn about optimizing its therapeutic impact. Currently, TMS is approved as an aid for cigarette smokers in some regions, but its effectiveness in long-term addiction treatment remains a subject of ongoing research. But that does not stop TMS clinics from using a regulatory loophole and promising to cure people from addiction for as much as $15,000 for a full therapy.
New Human Embryo Models Spark Needless Controversy
Recent news about the creation of synthetic human embryos revived the fears of human cloning or unethical experiments on embryos. But as Insoo Hyun argues in this article, there are safeguards against such practices. Firstly, it is forbidden to generate pregnancies using human embryo models in many countries. On top of that, there are legal requirements and processes needed to be met before any experiment on embryos can take place. And secondly, it is impossible for a synthetic embryo to be transferred into the uterus to establish a pregnancy. The embryo at this stage just does not have the supporting tissue to develop into a healthy pregnancy.
🧠 Artificial Intelligence
xAI
After first teasing the idea at the beginning of the year (when OpenAI and ChatGPT were on everyone’s minds) Elon Musk finally unveiled xAI, a new AI lab with a goal to “understand the true nature of the universe”. We don’t know yet what the new company will be exactly working on. Greg Yang, one of the founders of xAI, tweeted that the new company will be working on Ai that “will enable everyone to understand our mathematical universe in ways unimaginable before”. In the past, Musk was teasing the idea of TruthGPT, a competitor to ChatGPT. The team currently consists of people who previously worked at DeepMind, OpenAI, Google, Microsoft and Tesla, and will be led by Elon Musk himself.
ChatGPT's explosive growth shows first decline in traffic since launch
Reuters has reported a 9.7% decrease in traffic to the ChatGPT website in June compared to May, with a corresponding drop of 5.7% in unique visitors. This decline marks the first instance of reduced traffic for ChatGPT since its launch in November 2022. We have seen the first signs of lower interest a month ago when SimilarWeb reported a plateau in visits to ChatGPT.
Claude 2
Anthropic has released its newest large language model, Claude 2. According to Anthropic, the new model is smarter and better at taking exams, coding and solving math problems than the previous model. It’s also safer and less likely to provide a harmful or incorrect answer. Claude 2 has also a larger context window. It can take in now 200K tokens, which is equivalent to about 150,000 words (for comparison, GPT-4 allows a maximum of 8,192 tokens with some versions extending this to 32,768 tokens). In other words, it is possible to give Claude 2 an entire book to summarise (Anthropic says that initially, the context window will be limited to 100k tokens). Anthropic plans to offer Claude 2 via API to businesses to power their AI products.
GPTrolley
The Trolley Problem. A seemingly simple question that can ignite hours of ethical discussions. Thanks to GPTrolley, we can now use ChatGPT to give a definite answer what is the best answer to this dilemma. Simply input the options you wish to place on the tracks, and the AI will select one of them while providing an explanation for why it chose that particular option over the other.
Tired of telemarketers? Now you can turn the tables on them with this clever bot
Jolly Roger is one of the best AI apps released this year. It is an AI-powered bot that talks to telemarketers on your behalf and wastes their time. You can listen to it in action here. Jolly Roger is available in the US, UK, Australia, and New Zealand. Service is available to Canadians, but requires long-distance calling to the US.
🤖 Robotics
Humanoids ready to take first steps
Here is a good overview of companies working on humanoid robots, from well-known players such as Boston Dynamics and Agility Robotics to up-and-coming companies. If you are interested in this topic, check out my article on humanoid robots, too.
🧬 Biotechnology
▶️ Growing Rat Neurons... To Play Video Games? (27:03)
The Thought Emporium, Youtube’s chief bioengineer, returns with a project that involves growing rat neurons with the goal of making them play DOOM (I like how this meme is spreading beyond computer science). This episode focuses on growing neurons. The next episodes will focus on connecting neurons with electronics.
Synthetic Cells Stripped of Nearly All Their Genes Still Thrive Under Evolution
A recent study published in Nature has demonstrated that synthetic bacterial cells can thrive and evolve at a rate nearly 40% faster than their non-minimal counterparts. The researchers stripped down these artificial cells to a minimal genetic blueprint, which initially resulted in a loss of evolutionary fitness. However, over 2,000 generations, the streamlined cells regained their ability to survive, grow, and reproduce. These findings have significant implications, as they could lead to the development of synthetic bacteria that can produce life-saving medications, biofuels, or help with the bio-degradation of hazardous chemicals by harnessing the power of evolution.
A Protein Printer
This article from
H+ Weekly sheds light on the bleeding edge of technology and how advancements in AI, robotics, and biotech can usher in abundance, expand humanity's horizons, and redefine what it means to be human.
A big thank you to my paid subscribers and to my Patreons: whmr, Florian, dux, Eric and Andrew. Thank you for the support!
You can follow H+ Weekly on Twitter and on LinkedIn.
Thank you for reading and see you next Friday!