Profile avatar
scottleibrand.bsky.social
Built Open Artificial Pancreas (OpenAPS) with Dana Lewis. Day job at Netskope. DMs are open or email [email protected].
130 posts 316 followers 85 following
Prolific Poster
Conversation Starter

Nicholas Carlini wrote a super-popular post about working in software and security with LLMs ("Why I Use AI"); here's an almost equivalent post by a historian about using LLMs to do good history work: resobscura.substack.com/p/the-leadin...

inferencemagazine.substack.com/p/agi-is-an-...

I appreciate Marina speaking very openly about what a lot of us have felt (absolutely including myself) https://reason.com/2025/02/13/i-tried-to-fix-government-tech-for-years-im-fed-up/

KM3NeT xkcd.com/3053

These four points on DeepSeek seem very likely correct and important to understand about the economics of building AI models and what DeepSeek actually did, from the CEO of Anthropic. darioamodei.com/on-deepseek-...

1. It's Friday, which means FluView. Let's take a look. This has been a very full-on flu season — the worst since well before the #Covid pandemic. #CDC says 2024-25 is now classified as a severe flu season, the first since the very bad 2017-18 season. Red arrow this year, blue arrow 2017-18.

I generally agree with this important caveat, but… … I have now spoken with C-level leaders & lawyers, who assigned teams to check Deep Research. They found few flaws & fact checking took much less time (and less highly skilled work) than was saved by AI. That seems to be a critical tipping point.

Incoming Asteroid xkcd.com/3049

The maneuvering you are seeing now, both in Paris & among big AI players, is a sign that insiders increasingly believe the growth of AI capabilities will continue in the future towards AGI-ish systems without hitting a wall. Many other normal considerations are taking a back seat to this belief.

I can’t believe it - after years of advocacy, exclusionary zoning has ended in Cambridge. We just passed the single most comprehensive rezoning in the US—legalizing multifamily housing up to 6 stories citywide in a Paris style Here’s the details 🧵

For #SuperbOwl Sunday, here's the website @natbat.bsky.social and I built a few years ago to help you find your nearest superb owls www.owlsnearme.com

This is good news! "Right now, if you want to make something electricity-intensive, you go to China, with its ultra-cheap coal power. In two decades, there’s a good chance you’ll go to the Middle East instead, with its cheap abundant solar power."

VPP offerings are booming in Texas, with now Solrite & Sonnen partnership offering two 20 kWh batteries & solar panels for zero upfront payment. Customers lease equipment and pay 12 cents/kWh. VPPs cut gas burning & could cut peak demand by 10%-20%. www.canarymedia.com/articles/vir... #energysky

This is a fantastic bluesky feature that makes reading your timeline in long threads a lot cooler and easier. Turn it on and see what you think

Phrases never uttered: - "Without carried interest deduction I wouldn't have started my PE firm" - "Without Qualified Small Business Stock Exemption I wouldn't have started my tech firm" - "Without Roster Depreciation this pro sports team wouldn't exist" get rid of all this crap.

Our op-ed @nytopinion.nytimes.com addresses the surprising results of recent medical studies that showed A.I. alone outperformed physicians using A.I. With @rajpurkar.bsky.social Here is a summary Table, an overview, and a gift link to the op-ed erictopol.substack.com/p/when-docto...

What if you *could* do the things that you thought were impossible (or feel too hard, or you don't know where to get started)? Speaking of overwhelming, what do you do if the pace of AI development feels rapid and you're not sure where to start? Or what to do? Blog: bit.ly/3WPZYOl

One thing academics should take away from Deep Research is that a substantial number of your readers in the future will likely be AI agents. Is your paper available in an open repository? Are any charts and graphs described well in the text? Is it well-titled? Probably worth considering these…

The chance of asteroid 2024 YR4 hitting out planet in 2032 is now 1.5%, or 1 in 67. The OVERWHELMING likelihood is that the asteroid will miss Earth. But for the first time ever, we might have to seriously consider a deflection mission soon. Let me explain. 🧵 (1/x)

An acre of land can power a car to go 13,000 miles per year using biofuel (corn ethanol), or 900,000 miles per year using solar.

Stromatolites xkcd.com/3046

Thinking about how medical research is similar to a game of telephone, in terms of understanding how different information can be presented at the same time in different papers. (And helping patients understand how there might be conflicting information, and why) More thoughts: bit.ly/4jWGx0z

Two order of magnitude cost reduction in a month for reasoning models from OpenAI (also means large reduction in energy used to run the model)

I’ve found it surprisingly useful to have ChatGPT read me off (modified) recipe directions one step at a time. I made a GPT (chatgpt.com/g/g-678096a6...), and now I give it a recipe and whatever random extra instructions, and then put it into voice mode to walk me through it.

The new OpenAI model announced today is quite wild. It is essentially Google's Deep Research idea with multistep reasoning, web search, *and* the o3 model underneath (as far as I know). It sometimes takes a half hour to answer. Let me show you an example. 1/x

"It felt more like a supercomputer on wheels than any car you can buy in the United States right now." Chinese EV makers show off their advanced cars at CES in Vegas, making auto & tech fans drool and US automakers very nervous. insideevs.com/news/746628/... 🔌💡 🔌🚗

An asteroid first spotted in December has a 1.2% chance of hitting Earth in 2032. Scientists are tracking the space rock to learn more about its size and trajectory.

www.hyperdimensional.co/p/novus-ordo...

The Zeekr 7X, with the 800-volt Golden Battery, is capable of adding 1,270 miles of range per hour. insideevs.com/news/749125/...

Influenza is back up *again*. I’m (unpleasantly) surprised! Full report coming on Monday. caitlinrivers.substack.com

Call me risk averse (or maybe overly anxious would be a better description), but at a 1.2% probability I did click through to check the size of this asteroid (about the same or a bit bigger than the impactor that created meteor crater) 🧪⚒️🔭

D O N ' T P A N I C A decently big (50-meter) asteroid called 2024 YR4 has a 1.2% chance of hitting Earth in 2032. Here's what we know, and why I'm not terribly worried about it. badastronomy.beehiiv.com/p/will-the-a... 🧪 🔭

The people who are not particularly surprised by recent developments in AI are by and large saying AGI could be soon. The people confidently saying AGI won’t be soon are the same people who kept being very surprised by advancements over the past years.

A new feature in Carb Pilot - a widget! (Available today in 0.1.4: bit.ly/Carb-Pilot-iOS) Carb Pilot is an iOS app that allows you to custom track only the macronutrients you care about. Track any of carbs, fat, protein, and/or calories. Change your mind any time, & integrates with HealthKit.

'Bare Photovoltaic Tether...conductive boom in a magnetic field to provide a propulsive force...counteract 2km/month orbital drop from ISS w/15 km long tether, 97% covered in solar panels totaling 8.3 kW of power generation...hundreds of tons of hydrazine rocket fuel saved over multiple decades'