Alec Newcomb

Serial Troublemaker. Helping 12,000+ marketers actually use AI.

[Alec Newcomb @alecnewcomb]

My day is toast as I installed omarchy by @ddh and it's AMAZING for running your agent fleet. I'm going to swap all my windows machines.

[Alec Newcomb @alecnewcomb]

@shawngorham If your a contractor ChatGPT via voice mode is the way to go get a pair of Meta Glasses or mic setup you like and talk to it in loops. https://t.co/VwqUrbzhR9

[Alec Newcomb @alecnewcomb]

@jimmykim @EcomMikeGalvin Save yourself the pain > beehiiv. I lost a year I'll never get back to Kit.

[ClaudeDevs @ClaudeDevs]

Claude Code artifacts can now call MCP connectors, letting you build dashboards and apps that can fetch information and take actions for each viewer on demand.

Available on Pro, Max, Team, and Enterprise plans. Not available on publicly-shared artifacts.

[Voices of WW2 @VoicesofWW2]

On this day in 1944, Theodore Roosevelt Jr. died in his sleep in a stone farmhouse in Normandy. He was 56 years old, and he had spent almost his entire adult life trying to be worthy of a famous last name.

He was the eldest son of President Theodore Roosevelt. In the First World War he went to France and was gassed and badly wounded at Soissons leading his men. That same summer his younger brother Quentin, a pilot, was shot down and killed over France. Ted came home with lungs and a leg that never fully recovered, and before he even left Europe he helped found the American Legion so that ordinary soldiers would have someone looking out for them.

Between the wars he did almost everything. Governor of Puerto Rico. Governor General of the Philippines. Businessman, explorer, writer. He could have spent the Second World War safe behind a desk. Instead, at 54, arthritic and walking with a cane, he talked his way back into uniform and into combat.

By 1943 he was fighting in North Africa and Sicily under Terry Allen, and their loose, unpolished, soldier-first style rubbed General Patton the wrong way. Patton had them both relieved of command. Roosevelt didn't sulk. He asked for another job, any job, as long as it kept him near the fighting. They made him assistant commander of the 4th Infantry Division.

Then came D-Day. He hid a heart condition from the Army doctors. He wrote to his commander three separate times, in writing, begging to go in with the very first wave rather than watch from a ship. He was the only general to land in the first wave on any beach that morning, the oldest man in the invasion, walking through machine gun fire with a cane in one hand and a pistol in the other.

The boats came in a mile off course. Officers froze. Roosevelt limped up and down the beach under fire, studied the ground, and said, "We'll start the war from right here." Then he spent the morning waving men forward and sorting out the chaos so calmly that terrified 20 year olds looked at this old man with a cane and decided that if he wasn't scared, they wouldn't be either.

His son Quentin, named for the uncle killed in the last war, landed at Omaha Beach the same morning. They were the only father and son to come ashore together on D-Day.

He died a month later. A heart attack in his sleep. And here is the part that gets me. On the very day he died, the orders had just come through promoting him to major general and giving him his own division. He never saw the paperwork. He never knew he'd earned the Medal of Honor either.

At his funeral his pallbearers were seven of the most famous generals of the war, Bradley, Hodges, Collins, Barton, Huebner, and George Patton. The same Patton who had fired him. Patton wrote in his diary that Roosevelt was one of the bravest men he had ever known.

Years later Omar Bradley was asked to name the single most heroic thing he witnessed in all of World War II. He didn't pause. He said, "Ted Roosevelt on Utah Beach."

[andrew chen @andrewchen]

playing with local AI models and doing some research and came across a fun extrapolation:

"consumer grade graphics cards will be running Fable-equivalent models by 2029"

the argument:

1) been said that LLMs are really a lossy compression process for squeezing humanity's knowledge down into a chat box. Roughly you could say the entire internet/books/media/whatever is about ~200-300T tokens which gets "compressed" down to ~1T params of frontier models, so imagine a 400:1 or so compression in bytes. (We have a lot more video and world data, we'll discuss that later...)

2) People further quantize the models down (seemingly at decent quality) from 16-bit to 4-bit (NVFP4 ftw) and it seems pretty good. So that's more like 1,600:1 or so. Research suggests models only store ~2 bits of knowledge per parameter anyway — the weights were mostly air, which is why quantization works at all.

3) furthermore, we seem to be getting better at this compression, whether it's with MoE, pruning, etc. The Qwen 27B dense model now is equivalent to higher parameter models from a few years ago. 1,600:1 today and 1,600:1 in 5 years will have completely different results

4) I was looking to see if there's a Moore's Law thing happening here, and it's been measured: the "Densing Law" found capability-per-parameter doubles every ~3.3 months. Not sure how well it'll hold, but it says somethign like: "Every 3 months, the size of model needed to represent humanity's knowledge drops by half."

5) Not sure this holds though bc presumably, there's some kind of asymptote. Won't compress down to zero, the same way that modern image/video compression has theoretic limits too

5) Video is a zillion frames, almost zero semantic density. 99.99% of every frame is stuff a physics prior already predicts. World models will post compression ratios in the millions-to-one. But, nevertheless, there will be a ton of new facts seen simply by observing, that was never written down. But even without dealing with all this, today's text-oriented frontier LLMs are already pretty amazing

6) So the crazy idea here is that ultimately irreducible kernel representing humanity's knowledge might compress down to... tens of GB? May be small enough to fit onto a consumer grade GPU. Today, a consumer grade GPU for playing video games might have 32GB of memory on it, but the 27B parameter model that thing can run is getting smarter and smarter each year. Will the equivalent of Fable be able to run on a high-end consumer GPU in a few years?

This sounds crazy, but GPT-4 was rumored ~1.8T params, needed a rack of A100s, cost tens of millions to train. Today's open weight models with 27B parameters can do that with better data, distillation, and architecture.

That's the big question. Densing Law says "consumer GPUs might be running Fable-equivalent in 2028" and it might be possible, and even though that would be 100x? Seems nuts

[Ben 🚜 @2022Benjii]

Whoever made this

Take a fecking bow

This is class

[Marc Andreessen 🇺🇸 @pmarca]

Interesting.

[Luke Pierce @lukepierceops]

Starbucks spends $400 million a year on software. Yesterday they announced they're moving off IBM and Microsoft to build their own custom systems in-house.

IBM dropped 3% and Salesforce dropped 4% on the news.

And honestly this is, unequivocally, the biggest signal I've seen since OpenAI and Anthropic launched their consulting arms back in Q1. The largest companies in the world are done paying for software that half fits how they work.

We saw this coming about a year ago. Moved everything we build off Airtable and low-code tools and went fully custom. Already paying off, and it's only going to compound from here.

This is the opportunity right now.

You get all of a company's data into one system. You build out a single operating system for the entire business. You cut out bad, redundant processes. Then you layer AI on top of it, under the correct processes.

That's the core of AI consulting. Helping companies actually operate better.

There are a lot of fly-by-night offerings circulating right now when it comes to Ai Services.

For example, 'second brains'.

Throwing scattered data into a second brain while the processes underneath stay broken does nothing. The companies who will absolutely destroy their competition over the next 5 years are rebuilding how they work from the ground up.

Starbucks is showing you what other companies will be doing over the next several years.

Your job is to position yourself to facilitate that process for as many companies as you can.

[Alec Newcomb @alecnewcomb]

@kidpakerot warmr

[Dan Shipper @danshipper]

the state of the race between Anthropic and OpenAI:

- Ant: Make the biggest, most powerful, most expensive model possible—Fable. Use that to hit RSI faster and break away from the race.

- OpenAI: Make a powerful, useable model that you can serve efficiently / cheaply with a ton of compute—5.6 Sol. Focus on a ton of post-training rather than raw model size.

5.6 is definitely a better daily driver model for the vast majority of people / use cases today. However, there are significant compounding benefits to continuing to push the frontier

game on!

[Alec Newcomb @alecnewcomb]

@thedennis Hooks

[Alec Newcomb @alecnewcomb]

@businessbarista ai

[Paul Graham @paulg]

Man do I want Paraguay to lose. Not just because they deserve to, but because I don't want to have to watch any more of them.

[Chaz Wargnier ☄️ @DataChaz]

✅ Claude Code
✅ Claude Cowork
✅ Claude Design
✅ Claude Finance
✅ Claude Science
⬜ Claude HR
⬜ Claude Analytics
⬜ Claude Marketing
⬜ Claude Sales
⬜ Claude Legal
⬜ Claude Logistics
⬜ Claude R&D
⬜ Claude Procurement
⬜ Claude Accounting
⬜ Claude Engineering

[claire vo 🖤 @clairevo]

No jira.
No figma.
No PMs.

Just a CTO, designer, 3 engineers + a permazoom.
(and ofc claude code)

That's how @GustoHQ launched the new AI-first version of their app in < 10 weeks.

I loved sitting down with @edawerd about a completely new way of building, including how he
- vibecoded a vision of the product during a delayed layover
- gathered organic interest in the project, and didn't get his feelings hurt when his code got thrown out
- giving permission to break every rule made everyone faster

Plus, we get his thoughts on what this means for the other 1000 people in the eng / pm / design org.

Full ep on yt: https://t.co/bAyUpCkXI7

[Alec Newcomb @alecnewcomb]

@bradmillscan I wasted a lot of time on gbrain.

[unusual_whales @unusual_whales]

Nvidia, $NVDA, has announced a warm-water cooling system that it says can dramatically reduce the amount of water a data center uses, eliminating “pretty much all water usage” inside the data center.

[Gary Vaynerchuk @garyvee]

You know what's about to explode? Analog. The real world.

[Andrew Lynch @andrewglynch]

if a toddler puts a sticker on you, that’s a sign of great respect in their culture