Weekly Dose of Optimism #72
Gemini, Qubits, Brain Implants, saRNA, Biotech on Mars, Age of Miracles
Hi friends 👋,
Happy Friday and welcome back to our 72nd Weekly Dose of Optimism. We’re feeling optimistic over here at Not Boring HQ — maybe it’s the holiday spirit or maybe it’s the constant drip of world-changing, positive news we keep on getting to share with you each week. Either way, we’ll take it.
Let’s get to it.
The Weekly Dose is brought to you by…Alto
Alto allows individuals to invest in alternative assets with their retirement funds through a self-directed IRA. It’s perfect for long-term investors in crypto, fine art, real estate, private credit that want to significantly reduce their tax burden in the future. Here’s an example of how investing with a Roth IRA can reduce your future tax burden:
Let's say you make a $10,000 investment in an asset with a Roth IRA. By the time you're 59 and a half, that investment has increased in value to $100,000.Â
If you invested with non-qualified cash, you could pay up to 20% in capital gains taxes over the lifetime of this investment, meaning up to $18,000 of potential gains is lost to taxes. The Roth IRA allows you to potentially avoid the entire tax liability, meaning $0 on the gains is lost to taxes. Of course, IRA rules and regulations apply, and you should seek advice from a tax professional when making investments.
Alto makes investing in alts with a Roth IRA easy. Even better, Alto just launched Alto Marketplace, a capital raise platform that connects individual accredited investors to leading funds and exclusive opportunities.
(1) Introducing Gemini: our largest and most capable AI model
From Google
Gemini is the result of large-scale collaborative efforts by teams across Google, including our colleagues at Google Research. It was built from the ground up to be multimodal, which means it can generalize and seamlessly understand, operate across and combine different types of information including text, code, audio, image and video.
Google launched Gemini, its newest and most advanced AI model yet. The model is natively multimodal, meaning it can pretty seamlessly operate against text, code, audio, image and video and it’s highly performant, beating GPT4 on both text and multimodal performance benchmarks.
(To be fair, as others have noted, it uses a different method than GPT-4 in the benchmark (CoT32 vs. 5-shot, respectively, and the video is not as one-shot as it seems and relied on more sophisticated prompts than the ones shown in the video.)
But still! In Gemini, we have the (pretty advanced!) beginnings of an AI model that you could imagine weaving its way into how we go about daily life. It’s not just a text summarizer or image identifier (although it does those things better than any model in the world.) It’s a cohesive model that you could easily imagine enhancing all of Google’s existing core products, but, perhaps more importantly, powering a whole new suite of more immersive future products. Perhaps all Google Glass needed was Gemini.
From an industry perspective, this was a pretty strong move from Google. It wasn’t around some planned Demo Day, nor was it an overly reactive response to an OpenAI breakthrough. In my opinion, Google frankly didn’t need to launch Gemini yet. But it did — reminding all of its competitors that they’re still Google and, likely, the leader in AI. This will, of course, prompt some kind of response from OpenAI or Meta or Anthropic over the next couple of weeks, which will in turn pull up some previously scheduled advancement from Google. This is acceleration.
One fun fact about Gemini: Google co-founder Sergei Brin was a core contributor on the technical paper behind Gemini. My man is worth over $100B and still cooking in the lab! (h/t @linasbeliunas for the find)
(2) DARPA-Funded Research Leads to Quantum Computing Breakthrough
From DARPA
A team of researchers working on DARPA’s Optimization with Noisy Intermediate-Scale Quantum devices (ONISQ) program has created the first-ever quantum circuit with logical quantum bits (qubits), a key discovery that could accelerate fault-tolerant quantum computing and revolutionize concepts for designing quantum computer processors.
DARPA researchers have developed quantum circuits that utilize logical quantum bits (qubits), which are essential for fault-tolerant quantum computing. Logical qubits differ from physical qubits in that they are error-corrected and can maintain their quantum state more reliably.
Quantum computing is generally error-prone due to the delicate nature of qubits. Unlike the classical bits we use in regular computers, which are either 0 or 1, qubits can exist in multiple states at once. This makes them both incredibly powerful for certain computations and very sensitive to their environment. On Moment of Zen, Guillaume Verdon/BasedBeffJezos said that error correction uses up ~99.9999% of the computing resources in a quantum computer.
Logical qubits help solve this error issue by grouping together multiple physical qubits to create a more stable and error-resistant unit, allowing for corrections of errors that occur in individual qubits.
Creating logical qubits is a big step in quantum computing because it not only addresses that fundamental challenge of error correction but also paves the way for more efficient and scalable quantum computing architectures — so we can tackle the really big problems!
(3) Thalamic deep brain stimulation in traumatic brain injury: a phase 1, randomized feasibility study
From Nature
Six participants with msTBI, who were between 3 and 18 years post-injury, underwent surgery with electrode placement guided by imaging and subject-specific biophysical modeling to predict activation of the CL/DTTm tract. The primary efficacy measure was improvement in executive control indexed by processing speed on part B of the trail-making test…Processing speed on part B of the trail-making test improved 15% to 52% from baseline, exceeding the 10% benchmark for improvement in all five cases.
A promising study showed that electrode brain implants improved cognitive function on a group of five people that suffered from traumatic brain injury. Over 5 million people in the US alone are permanently disabled from a traumatic brain injury and there hasn’t been much in the way of promising therapies.
Dr. Schiff and his team identified the central lateral nucleus as a key hub for attention and cognitive functions and hypothesized that stimulating this area with electronodes could help improve cognitive function. The sample size (only 6 participants) was small, but it appears Dr. Schiff’s hypothesis is correct. It’ll be exciting to watch companies like Neuralink (and its competitors) race to commercialize the research behind these types of studies over the coming years, to bring a bit of relief to folks suffering from brain damage.
(4) Self-copying RNA vaccine wins first full approval: what’s next?
Elie Dolgin for Nature
The approval of yet another RNA-based vaccine for COVID-19 might not seem momentous. But the endorsement last week by Japanese authorities of a jab against SARS-CoV-2 constructed using a form of RNA that can make copies of itself inside cells — the first ‘self-amplifying’ RNA (saRNA) granted full regulatory approval anywhere in the world — marks a pivotal advance.
Much has been written about mRNA’s 20 year path to real-world applicability during the Covid pandemic. But that’s likely just the beginning of RNA’s story when it comes to how to develop medicines and treat diseases. ARCT-154, a covid vaccine recently approved in Japan, became the first 'self-amplifying' RNA (saRNA) drug ever approved. saRNA technology represents a significant advancement over conventional mRNA vaccines, offering potentially stronger and longer-lasting immune responses with lower doses. For example, ARCT-154 requires one-tenth to one-sixth as much vaccine per person as other RNA-based COVID-19 booster jabs — leading to lower production costs and less side effects.
Yes, the first vaccine is a Covid vaccine, but the technology opens the door to more versatile and efficient treatments for everything from the common flu to cancer.
(5) Colonizing Mars isn't a rocket problem
Elliot Hershberg for Century of Bio
As rockets get cheaper and bigger, the bottleneck for space colonization will be our ability to support biological life outside of Earth’s atmosphere.
Our friend and Not Boring Biotech Partner Elliot Hershberg explores the landscape of biotech-enabled sustainable space exploration. He breaks down ideas like:
Microbial Solutions for Space Habitats
Biotech-based Food Production
Medicine Manufacturing in Space
Materials and Construction
and provided a techno-economic analysis for space biomanufacturing. In Elliot’s view, colonizing Mars and deep space exploration is as much as biotech problem as it is a rocket problem. Seems like the rocket problem is well on its way to being covered — thanks Elon — but we’ll need a whole new class of optimistic biotechnologists to make sure we know how to biologically survive and thrive once we get to Mars and beyond.
BONUS: Age of Miracles, Episode 8: The Fusion Marathon
Packy here. There’s a joke that fusion is always 30 years away. That joke is no longer true. It’s going to be here within a decade. While we still don’t have commercial fusion after 80 years of research and development, the main metric in fusion - the triple-product - has actually been improving faster than Moore’s Law since the 1970s. It’s grown by five orders of magnitude!
In this episode, we get you ready for the coming fusion boom by explaining how fusion works, exploring the different approaches physicists have pursued since the 1950s, and cover the history of the multi-decade international race to create energy the same way stars do.
You can listen on Spotify, Apple, YouTube, or wherever you listen to podcasts. Enjoy!
P.S. A couple of weeks ago, we highlighted Bloomberg’s reporting that the US was leading a group of nations in a commitment to triple nuclear capacity by 2050 at the COP28 Climate Summit. It happened! More than 20 countries launched the Declaration to Triple Nuclear Energy.
The Declaration covers a bunch of the things we’ve discussed on Age of Miracles:
Mobilize investment in nuclear through innovative financing options.
Invite World Bank, international financial institutions, and regional banks to include nuclear in their lending policies.
Commit to small modular reactors and advanced reactors for power generation and wider industrial applications for decarbonization.
Support the development of nuclear for hydrogen or synthetic fuels production.
Promote resilient supply chains, including for fuel.
Big win for nuclear, and if they follow through, for the world.
That’s all for this week. If you have some time this weekend, check out Alto.
We’ll be back in your inbox on Tuesday.
Thanks for reading,
Dan + Packy
Great edition! 🔥