Love your work, love those who love you..live an adventure! Entrepreneurial. GovTech. Software, ideas, strategy, investing. #Dynamics365#MicrosoftPowerPlatform
Jeff Bezos: "If people want me to pay more billions, then let's have that debate, but don't pretend that that's gonna solve the problem. You could double the taxes I pay, and it's not gonna help that teacher in Queens.... Airbnb isn't causing high rents. What's really causing high rent is government intervention."
@EricSpracklen What’s wrong with it being 90 years old? I bought a 102 year old home and eventually sold it. Homes get updated, remodeled, fixed, etc. The structure itself was sound, the interior/plumbing/electrical/hvac had all been modernized. It was a modern home on the inside.
My longstanding theory is America’s lack of a single metropol, and our aversion to public transportation/density, is a key to our success & helped us avoid the fall of the rest of the Anglosphere
Urban and dense rowhouse suburbs always breed socialism; culturally & economically
@DonMiami3 How much work did it accomplish for $65 and how much would it have cost to have an analyst do that work?
Also, I have to keep the analyst busy with useful tasks full time to maximize value. With Claude if it’s a slow afternoon I just don’t pay it.
Huh? I love walking, I just much more enjoy walking through the woods then on city streets. Hence, I own a house that backs up to wooded park land and has a stream and paved trail.
I don’t think companies are going to commit themselves to a model or an inference provider because they like the harness. Harnesses will be a business but it won’t have massive lock-in, so it’ll be a competitive software business like many others - profitable but not easy. Inference will be a competitive commodity business. And models will have to compete on quality/speed to drive demand to get licensing revenue. Model routing will be a competitive business (SaaS and software) to get each prompt to the best model and inference provider. AI will be a big business but it’ll actually look like other competitive technical businesses.
All the discussion of home prices and boomers is just missing the mark. It isn't about what people bought houses for in the 70s and 80s - because it was expensive then. It got cheaper in the 90s and kept getting cheaper until it skyrocketed starting in 2022. The anger isn't that boomers got a deal on their first home, it's that a lot of 30-somethings didn't jump in when they should have over the last 10 years and are now are facing a much more expensive market (but still cheaper than someone buying a home in the 80s.)
https://t.co/dquyVHAnMz
Looking at home price is the wrong way. You have to look at affordability as percentage of income. Pre-1992 was not a good time. From 1992-2021 we were below 38% of median household income to pay monthly P&I on the median sale price of a home.
But what really has impacted perceptions is that from 2008 to 2021 we were 31% or less (often below 30%) and then in 2022 we jumped to 38% and have been there since. It has nothing do with boomers buying their houses in the 70s and 80s.
https://t.co/dquyVHzPX1
@JasnTru I really don't even get how this became such a discussion. It's largely just people's personal preference of how they prefer to live and do things in their life.
A year ago at GTC, Jensen brought out a DGX Spark in one hand and a MacBook in the other.
Yesterday, at GTC Taipei, Jensen brought out NVIDIA's new RTX Spark laptop in both hands.
This is the start of a new era of personal computing - the personal AI era.
In the new era, there are two competing platforms:
- @apple with macOS / MLX
- @nvidia with Windows / CUDA
Everyone will have an always-on personal agent that runs locally, constantly looking out for you, working for you proactively, monitoring the internet and talking to other agents. This will be a personal AI agent you own, that's private, that's aligned with you (not OpenAI or Anthropic). @karpathy calls it personal computing v2.
Let's set the scene for the new era of personal computing by diving into the one thing that will matter the most - the hardware.
The best hardware for local AI isn't what's running in a data center. It's a radically different problem. Here's a breakdown of the 3 most important things:
1. Memory.
LLMs are big. To run a model locally, you need to fit the entire model into memory. Apple (with Apple Silicon) and NVIDIA (with DGX Spark + RTX Spark) have both moved towards unified memory, which puts all the memory on one chip - leveraging cheaper LPDDR5X memory - useful for making more memory accessible to the GPU. The alternative competing architecture is a disaggregated CPU/GPU architecture - which is what the DGX Station uses. It has a large pool of slow LPDDR5X CPU memory (496GB @ 396GB/s), and a small pool of high-speed HBM3e GPU memory (252GB @ 7.1TB/s). It has a high bandwidth link (900GB/s) between the CPU memory and GPU memory, enabling fast disaggregated inference e.g. Attention on GPU, FFN on CPU. This enables running really large models like Kimi K2.6 (1T parameters) by offloading experts from CPU memory to GPU memory as they are needed. You could imagine something like this in a smaller form factor.
Hardware today:
- Apple M5 Max MacBook Pro: 128GB unified memory.
- NVIDIA DGX Spark / RTX Spark: 128GB unified memory.
2. Memory bandwidth.
In a data center, multiple user's requests can be batched together, which amortizes the cost of moving model weights into memory across many requests, pushing up arithmetic intensity to compute bound territory - meaning FLOPS matters a lot. Locally, everything runs at low batch size, which is low arithmetic intensity, i.e. memory bound - so FLOPS don't matter. What matters memory bandwidth. High memory bandwidth -> fast TPS. Low memory bandwidth -> slow TPS.
Hardware today:
- Apple M5 Max MacBook Pro: 617GB/s memory bandwidth.
- NVIDIA DGX Spark: 273GB/s memory bandwidth.
- NVIDIA RTX Spark: TBC.
3. Power.
In a data center, we talk about MegaWatts. Locally, we talk about Watts. Laptops have limited battery life. The best laptop batteries have a capacity of ~100Wh. LLM inference on a MacBook Pro consumes ~140W, meaning battery life with a persistent personal agent is less than an hour. This is unusable. The game will become how long can you run a useful agent on a laptop battery. Apple and NVIDIA will compete on how long an agent can run on battery - this will become the new battery life metric. This could be where an NPU or NPU/GPU hybrid really shines. Apple ANE has about 10x better power efficiency than the GPU on Apple Silicon (but has ~4-5x less memory bandwidth, with about the same FLOPS as the GPU). There will be an entire design space of how to build energy efficient agents - this will involve co-optimizing the harness, models, inference engines together.
Hardware today:
- Apple M5 Max MacBook Pro: Consumes 140W, battery capacity ~100Wh
- NVIDIA DGX Spark: Rated for 240W, consumes 140W. No battery (direct PSU).
- NVIDIA RTX Spark: TBC.
The hardware battle will be fierce, and I expect a move towards co-design, i.e. hardware designed *with* personal agent workloads. On top of this, models are improving, we're getting more intelligence per bit/watt, and open-source harnesses like @NousResearch Hermes / OpenClaw are improving rapidly. Within the next 2 years, we'll inevitably have unmetered, private Opus-4.8 / GPT-5.5 level intelligence running locally on a future version of a MacBook or RTX Spark. I like this future a lot better than the one where OpenAI / Anthropic control the intelligence layer of the internet and can rent-seek on intelligence.
Beyond this, NVIDIA is ahead on general AI ecosystem, i.e. the CUDA moat. Apple is ahead on local AI ecosystem, i.e. models quantized/rightsized for MacBooks, native macOS apps, and ease of setup. We'll see how this might change as the new RTX Spark also brings full native CUDA to Windows-on-Arm laptops for the first time, potentially closing the gap.
There are many other factors I haven't mentioned here, but I believe I've covered the timeless, most important things for the new era of personal computing.
Two books, “The Sociopath Next Door” and “The Myth of the Out of Character Crime”, should be required reading. Very clearly there are people who are just bad people - it isn’t “explained” by anything, we may search of excuses and find some but it’s not really causative.
[I will note that lead may have been a cause of some violent crime, but almost everyone with serious lead exposure has long aged out of being a factor.]
The fetishization of most things as “ideal” needs to stop. Everyone needs to figure out what is most important to their life and make tradeoffs. Where most of this dialog comes from is people not making tradeoffs, expecting they can have the home they want at a price they can afford in the community they want with the location they want - without recognizing that that isn’t how life works.
And I’m guessing in many of these startup cases it’s a handful of single twenty-something’s who practically sleep at the office but it’s also where they socialize and eat and play video games too. So you have people who basically go to the gym for a workout (maybe) and a shower (hopefully) and then sleep but are otherwise “at work”, but they aren’t coding 14 hours straight for a straight year or anything.
@ElvyraLei@RudeOnion@Lagerale Something isn’t adding up: rents are high but income is very low, so who is paying the rents? It’s possible you need to figure out a plan over the next five years to increase your income and relocate.
@BillMcGill666@mcj3379@RudeOnion And then one day your kids will graduate and you’ll be done tuition still living in a low cost home enjoying an amazing life…this is how it’s done!
@ElvyraLei@Lagerale@RudeOnion If where you live is too expensive to live on the best income you can achieve in that area you have to make some hard choices. You can stay there and struggle or you can relocate.
@MannOhneListe@xwanyex Nothing wrong with that, but it’s just a personal preference. When I was single living where I walked to the grocery store I still minimized trips.
The problem here is that there are actually two types of kids living in their parent’s basement:
1) Winners: building up massive savings, working hard, living smart and frugally for a few years. They’ll bust out and buy homes - because they’ve saved up a 5% down payment, have income security, and can spend time growing into a house payment b/c they are used to a frugal lifestyle. All of a sudden they’ll be in their forties, married, some kids, and have savings and equity and a good life.
2) Losers: not earning, not saving, not learning. Whining on X about how everything is wrong and it’s all other peoples fault. All of a sudden they’ll be in their forties, unmarried, no kids or romantic prospects, no savings or investments, and no path to a good life.
Kids can choose 1 or 2.
Kids are living in their parents’ basements because wages are stagnant, gas prices are skyrocketing, groceries are expensive, student loan debt is crushing, medical insurance is costly - if they can get it, and there is no affordable housing.
But sure, let’s teach them Dave Ramsey’s envelope method. That will fix it.