〽️

A kid from the Bay. Brand Manager @cloroxco Weekends: Entrepreneur, Scout & Angel Investor @lvlupvc, Advisor @AskTaxGPT

[〽️ @mandizzay]

I’m healing a fractured ankle and on my butt most of the day.

Inspired by @databricks #dais releases this week, spent all day Saturday upgrading to Unity Catalog. Quite impressed with the features, but there’s room for improvement that I’m excited to watch.

The result?

Genie provided me with a complex SQL query I turned into a pipeline that automatically feeds a heat map of texts received from our clients from 2023-2024 (and updating every hour)

A fun use of being crippled this weekend, it’s time for a beer 👊🏾

[Noah @NoahKingJr]

“AI will replace you”

Me who already replaced myself with Claude:

[〽️ @mandizzay]

Every piece of software UI should be custom and generated on edge devices in the next 5 years

The future of software is hyper personalized and never persistent

[Google DeepMind @GoogleDeepMind]

Watch how fast Gemini 3.1 Flash-Lite can generate websites. ⚡

This browser creates each page in real-time as you click, search, and navigate. Give it a try → https://t.co/h3W5o1wItY

[Nick Greenawalt @motionbynick]

day 12 of putting ads on my toilet until i make $1,000,000

featuring @Clorox #ad

[〽️ @mandizzay]

How do I tell my kids

“Dad’s first advertisement was on a toilet and it went viral”

Jokes aside, I was sitting on my throne last monday and saw @motionbynick’s first post. Immediately began navigating the behemoth that is a Fortune 500 CPG

All for #toiletculture I hope I made the world proud

[Nick Greenawalt @motionbynick]

day 12 of putting ads on my toilet until i make $1,000,000

featuring @Clorox #ad

[〽️ @mandizzay]

First post on AI I agree with 100%

I think we’re closer to personal OS AI systems running on a MacBook in everyone’s house than than he thinks though

[Andrej Karpathy @karpathy]

Very interested in what the coming era of highly bespoke software might look like.

Example from this morning - I've become a bit loosy goosy with my cardio recently so I decided to do a more srs, regimented experiment to try to lower my Resting Heart Rate from 50 -> 45, over experiment duration of 8 weeks. The primary way to do this is to aspire to a certain sum total minute goals in Zone 2 cardio and 1 HIIT/week.

1 hour later I vibe coded this super custom dashboard for this very specific experiment that shows me how I'm tracking. Claude had to reverse engineer the Woodway treadmill cloud API to pull raw data, process, filter, debug it and create a web UI frontend to track the experiment. It wasn't a fully smooth experience and I had to notice and ask to fix bugs e.g. it screwed up metric vs. imperial system units and it screwed up on the calendar matching up days to dates etc.

But I still feel like the overall direction is clear:
1) There will never be (and shouldn't be) a specific app on the app store for this kind of thing. I shouldn't have to look for, download and use some kind of a "Cardio experiment tracker", when this thing is ~300 lines of code that an LLM agent will give you in seconds. The idea of an "app store" of a long tail of discrete set of apps you choose from feels somehow wrong and outdated when LLM agents can improvise the app on the spot and just for you.
2) Second, the industry has to reconfigure into a set of services of sensors and actuators with agent native ergonomics. My Woodway treadmill is a sensor - it turns physical state into digital knowledge. It shouldn't maintain some human-readable frontend and my LLM agent shouldn't have to reverse engineer it, it should be an API/CLI easily usable by my agent. I'm a little bit disappointed (and my timelines are correspondingly slower) with how slowly this progression is happening in the industry overall. 99% of products/services still don't have an AI-native CLI yet. 99% of products/services maintain .html/.css docs like I won't immediately look for how to copy paste the whole thing to my agent to get something done. They give you a list of instructions on a webpage to open this or that url and click here or there to do a thing. In 2026. What am I a computer? You do it. Or have my agent do it.

So anyway today I am impressed that this random thing took 1 hour (it would have been ~10 hours 2 years ago). But what excites me more is thinking through how this really should have been 1 minute tops. What has to be in place so that it would be 1 minute? So that I could simply say "Hi can you help me track my cardio over the next 8 weeks", and after a very brief Q&A the app would be up. The AI would already have a lot personal context, it would gather the extra needed data, it would reference and search related skill libraries, and maintain all my little apps/automations.

TLDR the "app store" of a set of discrete apps that you choose from is an increasingly outdated concept all by itself. The future are services of AI-native sensors & actuators orchestrated via LLM glue into highly custom, ephemeral apps. It's just not here yet.

[Nick Greenawalt @motionbynick]

So @pinesolcleaners paid me $1,000 to make this ad:

[Oleksandr Veremeyenko @alex_verem]

🚨 RAG is broken and nobody's talking about it.

Stanford just exposed the fatal flaw killing every "AI that reads your docs" product.

It's called "Semantic Collapse", and it happens the moment your knowledge base hits critical mass.

Here's the brutal math (and why your RAG system is already dying):

[Nicholas Fabiano, MD @NTFabiano]

Friends brains synchronize over time.

This brain activity can be used to predict your friends' purchasing decisions.

[Anthropic @AnthropicAI]

THE WAY OF CODE, a project by @rickrubin in collaboration with Anthropic:

[amit @amitisinvesting]

WHY DONALD TRUMP WANTS THE MARKET TO CRASH

….in the short term ⬇️

This chart below sums up the reasoning behind what the current adminstration is doing and why it is having adverse effects on the market.

Kris @KrisPatel99 did a great job today explaining this more in depth on the market open & I think the thesis checks out:

1. We have $7T of debt we need to pay in the next 6 months…if we don’t pay it, we’ll have to refinance.

2. The Trump admin does NOT want to refinance at a 4%+ rate…the 10yr at one point this year was 4.8%.

3. How do you get the 10yr to come down? Markets need to show weakness in growth, DOGE has to be perceived as actually working, interest rates need to come down.

The way to do that is to create massive uncertainties — aka tariffs — which can slow down growth in the short term, get the bond market to start BUYING bonds ASAP because of how scared they are of touching stocks (causing yields to fall which is what we need to refinance the debt) and then that gives the Fed the authority to lower rates which continues to bring yields down.

So, although conventional wisdom says tariffs are inflationary and the 10yr should be spiking on more tariffs — it’s actually going down because its bringing so much uncertainly to equity markets that people are selling stocks and buying bonds!

Which is exactly what the Trump administration wants to happen in the short term in order to bring refinancing costs down.

Short term pain for long term gain?

[vittorio @IterIntellectus]

it's here!

cortical labs just launched the first commercial biological computer
human neurons directly integrated onto silicon chips. programmable and adaptive, living computation

synthetic biological intelligence
1/

[Kash from TaxGPT.com @ChKashifAli]

🚨 We are excited to announce TaxGPT's $4.6 million round, backed by @ycombinator @LAUNCH, Rebel Fund, Mangusta Capital, and other funds, angels, and family.

Psst, we are hiring for 10+ open positions

[3PickupMusicMan @3PickupMusicMan]

Some dads play catch with their kids, others trade licks.

[Avichal - Electric ϟ Capital @avichal]

Ironic that we got free AI from a hedge fund and $200/month AI from a nonprofit.

[〽️ @mandizzay]

Long Apple 🫡

[signüll @signulll]

some people didn’t get my apple post about execs smiling this morning, so let me break it down a bit. the r1 news is massive for apple—like, generationally massive. apple bet early on model commoditization. they didn’t blow capex on some me-too race, didn’t overreact, didn’t pivot into panic mode. they just played it steady, focused, controlled. they bet on *edge inference*. the models don’t live in the cloud; they live on the device. & that strategy is now unequivocally proven directionally accurate.

in a few years, your iphone, ipad, mac, homepod, whatever will likely be running an llm natively. it’ll be fast. private. agi-level good, potentially. apple will own that end-to-end experience: the ux, the model, the silicon, the hardware. & when everything is deeply optimized, that’s where apple thrives. no one else plays this game better.

when models commoditize, apple wins bc the value shifts from the model itself to how it’s deployed, integrated, & leveraged. this was google’s moonshot with pixel which would’ve left apple high & dry. in few years every apple device will ship with agi on board just like every tesla shipped will come with a “driver” on board.

[Chamath Palihapitiya @chamath]

This report is long but very good.

“With R1, DeepSeek essentially cracked one of the holy grails of AI: getting models to reason step-by-step without relying on massive supervised datasets. Their DeepSeek-R1-Zero experiment showed something remarkable: using pure reinforcement learning with carefully crafted reward functions, they managed to get models to develop sophisticated reasoning capabilities completely autonomously. This wasn't just about solving problems— the model organically learned to generate long chains of thought, self-verify its work, and allocate more computation time to harder problems.

The technical breakthrough here was their novel approach to reward modeling. Rather than using complex neural reward models that can lead to "reward hacking" (where the model finds bogus ways to boost their rewards that don't actually lead to better real-world model performance), they developed a clever rule-based system that combines accuracy rewards (verifying final answers) with format rewards (encouraging structured thinking). This simpler approach turned out to be more robust and scalable than the process-based reward models that others have tried.”

https://t.co/XRqgMNRzg1

[Jared Friedman @snowmaker]

Lots of hot takes on whether it's possible that DeepSeek made training 45x more efficient, but @doodlestein wrote a very clear explanation of how they did it. Once someone breaks it down, it's not hard to understand. Rough summary:

* Use 8 bit instead of 32 bit floating point numbers, which gives massive memory savings
* Compress the key-value indices which eat up much of the VRAM; they get 93% compression ratios
* Do multi-token prediction instead of single-token prediction which effectively doubles inference speed
* Mixture of Experts model decomposes a big model into small models that can run on consumer-grade GPUs

[Nicholas Fabiano, MD @NTFabiano]

The neural geometrodynamics framework of psychedelics.

🧵1/10

[anant sinha @siliconvmg]

Riley Walz explaining how he found where the United Healthcare CEO murderer probably fled to using Citi bike data

12/4/24

[〽️ @mandizzay]

If you ask me what’s stopping us from AGI? Inefficient compute.

Light (photon) processing versus electricity (electron) processing is the next obvious step change to unlock efficiency in neural networks.

Hybrid electron & light based GPU’s coming to you soon 👀

[Deedy @deedydas]

Using light as a neural network, as this viral video depicts, is actually closer than you think. In 5-10yrs, we could have matrix multiplications in constant time O(1) with 95% less energy. This is the next era of Moore's Law.

Let's talk about Silicon Photonics...

1/9