Weekly Dose of Optimism #124
Willow, Shipmas, Gemini 2.0, Trump Permitting Reform, BAMs
Today’s Weekly Dose is brought to you by … AlphaSense
AlphaSense recently acquired long-time Not Boring partner Tegus. Through their vast library of expert calls, the combined companies have more insight into how companies are actually using AI than anyone but the labs themselves. They’re where I turn when I’m trying to get the expert perspective on an industry or technology.
Next Tuesday, December 17th, I’ll be talking to AlphaSense’s Director of Research, AI, Sarah Hoffman, about generative AI’s remarkable advances in 2024 and how enterprises will move AI tools out of proof-of-concept and into full integrations in 2025. Sarah’s experience and position give her a unique perspective on what to expect and how your organization can best take advantage of AI.
Over 200 people are already signed up. Register today to secure your free spot.
Hi friends 👋,
Happy Friday and welcome back to our 124th Weekly Dose of Optimism. Just an extremely solid week over here at Optimism HQ. Google did what it does best: use its monopoly search profits to fund breakthroughs in other technological fields. OpenAI did what it does best: ship a ton of high-quality product quickly. Caltech did what is does best: put out important research. And Trump did what he does best: share out provocative policy statements on the internet that, behind the ALL CAPS, actually makes a fair bit of sense.
Let’s get to it.
(1) Meet Willow, our state-of-the-art quantum chip
From Google Quantum AI
Today I’m delighted to announce Willow, our latest quantum chip. Willow has state-of-the-art performance across a number of metrics, enabling two major achievements.
The first is that Willow can reduce errors exponentially as we scale up using more qubits. This cracks a key challenge in quantum error correction that the field has pursued for almost 30 years.
Second, Willow performed a standard benchmark computation in under five minutes that would take one of today’s fastest supercomputers 10 septillion (that is, 1025) years — a number that vastly exceeds the age of the Universe.
Big week over at Google. The company’s latest chip quantum chip, Willow, represents a major leap forward in the field of quantum computing. And, if we crack quantum computing…basically any problem that can be solved, will be solved. Cryptography. Materials science. Drug discovery. Point a quantum computer at these extremely complex topics, and a question that would take previous generations of supercomputers extremely (extremely) long periods of time to answer, can be in answered almost instantly.
There were three big takeaways from Google’s announcement of Willow.
(1) Willow effectively reduces error rates as the number of qubits increases, overcoming a major hurdle in quantum computing and paving the way for practical, large-scale quantum machines.
(2) Willow, can perform in under five minutes computations that would take today's fastest supercomputers 10 septillion years, showcasing unprecedented processing power.
(3) The founder of Google Quantum AI, Hartmut Neven, wrote that that Willow’s performance was so powerful that it “had to have 'borrowed' the computation from parallel universes.” The chip is so powerful that it implies the existence of the multiverse.
By the way, this is Hartmut Neven…
We’re not technical enough to understand whether Willow is the breakthrough that unlocks quantum, but we’re practiced enough vibe readers to suspect that this is the dude who’s going to make it happen.
We’re extremely excited to watch this space as quantum computing moves out of the lab and into practical real-world use cases. The Google team (and a number of other players) seem to be clipping off milestones at a quicker and quicker pace here, so we wouldn’t be surprised if quantum arrives to our sliver of the multiverse soon.
From OpenAI
It’s the most wonderful time of the year. Merry Shipmas.
We’re in the middle of 12 Days of OpenAI, the company’s showcase of its newest technologies and its overall product velocity. So far, they have not disappointed.
Day 1: o1 and ChatGPT Pro
o1: An advanced AI model with enhanced reasoning capabilities. ChatGPT Pro: A premium subscription offering unlimited access to advanced features.
Day 2: OpenAI's Reinforcement Fine-Tuning Research Program
Expanded alpha access for researchers and enterprises to fine-tune AI models on complex tasks.
Day 3: Sora
A video generation model enabling users to create and edit videos from text prompts.
Day 4: Canvas
An interface for collaborative writing and coding with ChatGPT, enhancing creative workflows.
Day 5: ChatGPT in Apple Intelligence
Integration of ChatGPT with Apple's AI ecosystem for seamless user experiences.
Day 6: Advanced voice with video & Santa mode
interactive voice and video communication with festive, personalized holiday-themed visual effects and responses.
When Sam said they’d steamroll you, I think this is what he meant…
We’re very excited to see what the next 6 days of Shipmas have in store…you gotta believe they have at least one more major release under the tree.
(3) Introducing Gemini 2.0: our new AI model for the agentic era
From Google
Not to be outdone by the quantum guys down the hall or the upstarts over at OpenAI, Google DeepMind released Gemini 2.0 on Thursday, its most powerful model yet. The release marks a significant step forward in the development of agentic AI, systems that can think, plan, and take action on behalf of users. Gemini 2.0 can understand context, generate creative text formats, translate languages, write different kinds of creative content, and answer questions informatively. The model is expected to power many of Google’s products, from search to Bard.
The coolest, most novel product, which we don’t have access to yet but can’t wait to try, is Deep Research, which makes research reports for any subject you wish by taking its time, checking a lot of sources, and pulling the results into something cohesive.
Based on replies and commentary on X, it seems like Deep Research is about as good as hiring a good college student to do research — unlikely to change the trajectory of your project, but a useful timesaver.
More generally, it’s just really, really cool to watch this AI arms race in real time. Every week there is some big, splashy advancement that pulls forward innovation across the industry. OpenAI launches new product? Cool. The next week Google launches a similar product, but faster. The next week Meta launches an open sourced version of that product. The following week Anthropic launches a more thoughtful version of the product. And then finally OpenAI (or any of the players) just totally start the cycle again with another big release. What’s happening in AI is the most legible, public display of innovation I’ve ever seen in my lifetime. That’s not only good for AI development, but I’m guessing it also pulls forward innovation in other industries — AI is just setting the example that you can move industries forward both thoughtfully and quickly.
(4) Trump Permitting Reforms
From Donald J. Trump on Truth Social
Speaking of watching things get built fast…
Earlier this week, President-elect Donald Trump “Truthed” his plans to expedite federal permits for $1 billion-plus investments in energy and infrastructure projects, including environmental approvals. That this policy comes from a former real estate developer, a class of entrepreneur that is particularly weary of permitting regulations and environmental approvals, is not exactly surprising. But in Trump’s America, IT’S TIME TO BUILD, and a policy like this could accelerate large-scale infrastructure and energy projects and attract substantial domestic and foreign investments.
To be clear, some people really don’t like this plan. They argue this violates NEPA and prioritizes corporate interests over community and environmental impacts. The critics are not exactly wrong — the plan not only likely violates NEPA, but is an outward challenge of it. Whether or not that’s a bad thing, I guess depends on your perspective on NEPA. We’ll see if and how the plan is actually implemented over the coming months. There’s going to be an interesting balance between regulatory oversight and economic growth over the next four years, but who doesn’t like some good ol’ fashioned common sense permitting reform?
(5) Minuscule Robots for Targeted Drug Delivery
From Caltech
Wham! Bam! Thank you, Ma’am!
Caltech researchers developed bioresorbable acoustic microrobots (BAMs)—tiny, hydrogel-based spheres capable of delivering drugs directly to targeted areas within the body. These microrobots can navigate bodily fluids, be directed precisely via external magnetic fields, and release their therapeutic payloads upon reaching specific sites.
The precise and targeted drug delivery, combined with the fact that the BAMs can safely degrade after use, should reduce side effects and improve treatment effectiveness for various diseases. And the early results are promising: in mouse studies, BAMs effectively reduced bladder tumor sizes. Targeted, precise, safe and effective drug delivery? Yeah, I’ll take that!
Packy here with a quick ask. We’re two weeks in to Boring News and having a blast. We added follow-ups on previous stories like Fed cuts, Mangione’s Manifesto, and TIME Person of the year. We’ll keep working to improve it. I’d love for this to be an easy five minutes to start your day. We may add the Weekly Dose Audio to the feed, too.
So the ask: go listen and subscribe on Spotify…
or YouTube…
And let us know how we can keep making it better.
Have a great weekend y’all.
Thanks to AlphaSense for sponsoring! We’ll be back in your inbox on Tuesday.
Thanks for reading,
Dan + Packy
I have working on a phd proposal for a few months and would have really loved to use Deep Research to save me some time. Mostly I have just used review papers (which seemed to have become much more common since 2018) and Claude for putting my disparate thoughts into coherent sentences.
Google's Deepmind will definitely be remembered as the Bell Labs of our age.
There’s going to be an interesting balance between regulatory oversight and economic growth over the next four years, but who doesn’t like some good ol’ fashioned common sense permitting reform
That Truth wasn’t permitting reform. Or a new policy. It’s just vibes. Let’s see if anything actually happens beyond it before we get excited about BUILDING.