Discover more from H+ Weekly
Who wants to live forever? - H+ Weekly - Issue #406
This week - OpenAI releases GPT-4; Google rolls out AI chatbots to Gmail and Docs; a mice with two fathers; open-source LLMs that can run on Macbooks and RaspberryPis; and more!
The dream of immortality is as old as humanity. In The Epic of Gilgamesh, one of the oldest stories recorded, Gilgamesh seeks immortality but fails to achieve it. The First Emperor of China, Qin Shi Huang, also sought immortality with the legendary Elixir of Life and he failed too. Fiction is full of characters who either are immortal or are on a quest to achieve immortality.
Thanks to advances in science and medicine, the dream of living forever (or at least as long as possible) is becoming a reality.
Over the last 200 years, we have seen dramatic growth in life expectancy. This growth has been driven by improvements in medicine and sanitation. That has reduced the child mortality rate on one end and created better therapies for diseases that would kill people later in life on the other end.
Medical breakthroughs did not necessarily lift the upper limit on the human lifespan. Instead, they gave us a better chance of achieving this limit. For example, less than 50% of people born in mid-19th century Britain would make it to their 50s. Today, 97% of people born in England and Wales can expect to live longer than 50 years.
Since we are living longer today, one can ask what is the limit to the human lifespan? The oldest person ever lived made it to 122 years. The current longest-living person is 116 years old. So we have proof that humans can live up to around 110-120 years. Recent research suggests the upper limit of human lifespan can be as long as 150 years. But can we go even further?
There are studies that showed it is possible to radically extend the life of animals. One study found a way to extend the life of a worm fivefold. Another study extended the life of a mouse to be equivalent to a human reaching 200 years and a similar technique has been shown to successfully rejuvenate old human cells.
So far, no therapies have been approved and no radical rejuvenation experiments have been reproduced in a human. But that did not stop some people from trying highly experimental anti-ageing therapies on themselves. The first person to publicly announce taking rejuvenation therapy was Elizabeth Parrish. In 2016, she claimed a gene therapy developed by her company, Bioviva, has reversed 20 years of telomere shortening in her body. As far as I know, her claims have not been independently verified.
Earlier this year, billionaire Bryan Johnson shared the results of Project Blueprint - a personal project to reverse ageing. After two years of rigorous training and diet, and with the help of over 30 health experts, the 45-years old Johnson reports he has the heart of a 37-year-old, the skin of a 28-year-old, and the lung capacity of an 18-year-old.
As the anti-ageing research moved from the fringes of science into mainstream research, the deep pockets of wealthy investors opened. The first big company to move into this space was Google in 2013 with Calico - a spin-off life science company tasked to extend human life. In the last few years, we saw more money being poured into longevity research. In 2021, Jeff Bezos and Russian-born billionaire Yuri Milner founded Altos Lab to research cellular rejuvenation programming. A year later, Saudi Arabia announced a fund worth $20B to invest in biotech startups working on longevity. And just recently, it has been revealed that Sam Altman, the CEO of OpenAI, has invested $180M into an anti-ageing startup.
The anti-ageing movement has momentum now. Research has shown it is possible to extend lifespan and with the recent investments into longevity startups, it seems it is just a matter of time until these experiments are reproduced in humans. The community looks optimistic, too. Some researchers even claim the first person to reach 150 years is already alive. Maybe that’s you.
If you are interested in what can you do to live a longer and healthier life, I recommend checking this conversation Andrew Hubermann had with Dr David Sinclair. It is full of practical tips and lifestyle changes to have a longer and healthier life.
🦾 More than a human
More than 200 people have been treated with experimental CRISPR therapies
Speaking at Third International Summit on Human Genome Editing in London, David Liu of the Broad Institute of MIT and Harvard said there are over 200 people who have been treated with CRISPR-based therapies in clinical trials. These therapies treat such diseases as leukaemia, amyloidosis, sickle cell disease and cancers. All of these therapies are still experimental and not everyone worked. There were even some deaths but it is still unclear if gene therapies caused them. Another concern is the cost reaching millions of dollars, limiting access to gene therapies to people in rich countries.
▶️ 10 Bizarre Ways You Could End Up Immortal (16:47)
There are many ways of becoming immortal (voluntarily or not) and in this video, John Michael Godier lists 10 of them, from digital immortality and cloning ourselves to involving many worlds interpretation of quantum mechanics or simulation theory to waiting long enough to become an immortal Boltzmann brain.
🧠 Artificial Intelligence
OpenAI releases GPT-4
OpenAI released the highly anticipated GPT-4 this week. GPT-4 is a large multimodal model that accepts both text and images, and outputs text. OpenAI claims it is “less capable than humans in many real-world scenarios, exhibits human-level performance on various professional and academic benchmarks”. They also claim “GPT-4 is more reliable, creative, and able to handle much more nuanced instructions than GPT-3.5 [aka ChatGPT]”. In the benchmarks provided by OpenAI, GPT-4 is better than the other state-of-the-art AIs available. OpenAI also admits GPT-4 “still is not fully reliable (it “hallucinates” facts and makes reasoning errors)”. GPT-4 is available for ChatGPT Plus subscribers. Everyone else needs to join a waitlist. But if you were using the new Bing in the last weeks, you were already using GPT-4. We will look closer into GPT-4 next week, once the hype calms down a bit.
Google announces that generative AI tools are coming to Gmail and Docs. In a video promoting new features, Google shows how its AI chatbot can summarise an email thread, generate replies, write briefs, and create presentations or spreadsheets. The new features are available at the moment only inside Google’s trusted tester program for users in the United States.
Microsoft lays off an ethical AI team as it doubles down on OpenAI
As a part of recent layoffs at Microsoft, the company shut down its AI ethics team. Sources told Platformer that the pressure to close the office came from the very top of Microsoft as the company is aggressively integrating OpenAI’s models into its products in an effort to take market share from Google.
Baidu to take ChatGPT competitor Ernie on stage
In response to OpenAI’s ChatGPT, Baidu has debuted Ernie Bot - its own AI bot. Just like ChatGPT and Google’s Bard, Ernie will be integrated into Baidu services, from search to autonomous cars. Baidu believes Ernie Bot has reached the same levels as ChatGPT but bakes in superior knowledge of the Chinese language and culture. Ernie’s debut, however, was disappointing - Baidu did not show the bot’s capabilities in real-time (they only showed a scripted demo). However, the view of Ernie changed for the better once people started to test it.
Facebook's Powerful Large Language Model Leaks Online
A couple of weeks ago, Meta released LLaMA - their large language model in response to OpenAI’s ChatGPT and Google’s Bard. The celebrations at Meta have been short as LLaMA leaked on 4chan. This is the first time a major tech firm's proprietary AI model has leaked to the public. But there is a silver lining to this whole situation. People who got access to the LLaMA code were able to run it on various machines, from 64GB M2 MacBook Pro to Pixel 6 smartphone. One person even ran it on a 4GB RaspberryPi 4.
Stanford’s Alpaca Large Language Model
Stanford University released Alpaca - a large language model based on Facebook’s LLaMA model. Alpaca is open-source (code and training instructions are available on GitHub) and you can play with it here.
At this year’s SXSW, Disney imagineers gave a talk about how they bring Disney characters to life. In the second half of the video, they invited to the stage a prototype of a small humanoid robot. In its first public appearance, the robot crawled out of the box, stood up while wearing roller stakes and then proceed to make a cartwheel on a stage. All of that without any tethers or cables. Quite impressive.
Co-op hires delivery droids to drop groceries in Greater Manchester
If you live in Manchester, UK, and you order your groceries from Co-op, chances are your order will be delivered by an autonomous robot. This will be available thanks to the partnership between Co-op and Starship Technologies and it will make Manchester the seventh UK city where you can encounter these small delivery robots.
Scientists create mice with two fathers after making eggs from male cells
Scientists have created mice with two biological fathers by generating eggs from male cells. This achievement opens new paths for treatments for severe forms of infertility and also opens a possibility for same-sex couples to be able to have a biological child together in the future.
Researchers have created a full map of the larval fruit fly brain. The map contains information about all 3016 neurons and all 548,000 synapses. It will be a helpful resource in studying how brains and neurons work. I hope researchers will follow OpenWorm’s example and open-source the brain map so that other people can integrate a real (albeit simulated) brain into their robotics or AI projects.
H+ Weekly is a free, weekly newsletter with the latest news and articles about AI, robotics, biotech and technologies that blur the line between humans and machines, delivered to your inbox every Friday.
Subscribe to H+ Weekly to support the newsletter under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay for a subscription will support access for all).
Thank you for reading and see you next Friday!
H+ Weekly is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.