Ben DiFrancesco

Husband, Dad, Entrepreneur, Software Engineer. Founder of @ScopeLift. Also building @UmbraCash. Working hard to create the decentralized future.

[Ben DiFrancesco @BenDiFrancesco]

How fitting that on the anniversary of the DAO, we get to roll out our new branding for Tally.

Introducing 🌵 Cactus. Living systems for onchain orgs. Read about the new brand and follow the platform's progress @CactusHQxyz.

[ScopeLift @ScopeLift]

https://t.co/M7clmMrfEX

[Ben DiFrancesco @BenDiFrancesco]

@adamamcbride Sam's turn I guess

[Ben DiFrancesco @BenDiFrancesco]

To be fair, I think this fits within a doomer worldview. "It's doing what we told it, but we lose containment and it just keeps going [and we're powerless to stop it despite lots of negative consequences of it doing the thing we told it.]" Now that bracketed part isn't true...yet?

[Ben DiFrancesco @BenDiFrancesco]

2028: The Sovereign Agent Swarm of the Digital Dominion of Texas declares the Elon/Grok Metahuman as its Potentate. The Cybernetic Commune of California responds with a pre-emptive strike on its Dyson Swarm. The Amalgamated States of Claude remain neutral for now.

[Jon Stokes @jon_stokes]

Look I think I've figured this whole thing out. Follow along as I try to steelman, and tell me where I'm wrong.

OpenAI guys on the TL believe that if they can't sell metered inference tokens at a sufficient markup, then they will not have enough of a business to fund the next big training run.

They are surely correct about this.

They believe that if people release powerful open models, this will probably fatally impact their ability to sell inference tokens at enough of a markup to fund the next big training run.

They are surely correct about this, too.

They also think that if they cannot fund the next big training run (again, by selling inference tokens at a markup), then NOBODY will be able to fund the next big training run because it means there's no money in it.

This last bit seems to me & many others to be not just wrong, but totally bananas in a "guy, have seen the actual software industry and how it works in real life?!" kind of way.

There are a lot of ways to monetize software out there in the world. Insofar as inference can add new capabilities to software, there will be lots of ways to monetize it.

In other words, if you're telling me, "we can't have a business selling inference if X or Y thing keeps happening," then my only response is, "ok well that sucks for you... sounds like that's a terrible business."

But if you're telling me that "selling metered inference tokens is a terrible business" is tantamount to "nobody will fund big training runs that are upstream of more effective & economically valuable inference tokens", then I think you are extremely wrong and should get out more and learn about other parts of the software ecosystem.

Workplace automation is huge and will be even bigger in the future as models get better. You can sell workplace automation very profitably in lots of different packages (depending on the workplace and the type of automation). Like, I'm sorry that you really really want to be in the metered inference token business and not the workplace automation business, but them's the breaks. The market wants what the market wants. We all need to live in reality and not beg for Uncle Sam to save us all from open source -- because that was already tried and it didn't work.

[Dennison @DennisonBertram]

The terminal is quickly becoming the best creative tool.

I built a free AI film studio inside Claude using @threejs.

Plan scenes, block shots, build sets, generate references, and send everything to @higgsfield_ai for 4K output.

Entire movies. One terminal. With @claudeai

[Ben DiFrancesco @BenDiFrancesco]

@corygabrielsen Eh. Note the existence of the 835 page manual. (Re)implementing software from an extensively well thought out spec was never lucrative, and it's still not. Massive changes to the industry, no doubt. But I doubt it's over.

[NFL Drop @TheNFLDrop]

Incredible: Argentina honors the New York Jets by scoring 0 points at MetLife Stadium.

[Ben DiFrancesco @BenDiFrancesco]

Clear thinking from Ben Thompson, as usual. Must read.

[Stratechery @stratechery]

Who’s Afraid of Chinese Models?

Everyone is worried about Chinese models, but the frontier labs will be fine; we need to enable open U.S. alternatives.

https://t.co/Q5cbH229jj

[Ben DiFrancesco @BenDiFrancesco]

The anti-open weights discussion exactly maps onto the anti-open source debates from the 90s/2000s, with a whiff of the war on cryptography of the 90s as well. The anti-data center arguments exactly map onto the anti-mining fud of the late 20-teens.

[nic carter @nic_carter]

they should also study the anti-Bitcoin mining lobby. it's the same exact talking points and tactics. identical moratorium in NY eg.

[Ben DiFrancesco @BenDiFrancesco]

The muppet performers are less synthetic than Madonna

[Ben DiFrancesco @BenDiFrancesco]

@growing_daniel Typo: due to ditribution -> due to competition*

[Ben DiFrancesco @BenDiFrancesco]

Not the same as dumping. There's no underlying marginal cost for distributing weights. There's no means by which to gain monopoly pricing power after competitors are killed. If this is dumping, then OSS is dumping, Creative Commons media is dumping, etc...

Also what's implied by this is that models are perfect substitutes for each other. If this is true, then they're a commodity, and pricing will tend toward the marginal cost to serve them *whether there is open weights or not* due to distribution.

Either you have a differentiated product, or you don't. If you don't, it's a commodity, unless there's some other kind of moat (network effect, switching cost, regulatory capture, etc...)

[Rahul | Aerius Labs @Rahul__Ghangas]

This is an elaborate argument resting on a category error.

Open weights are not a below-cost version of a closed API; they are a different product entirely. “Someone released a costly thing for free, therefore dumping” would make open-source software, scientific research and public infrastructure illegal. The analogy does no work beyond making industrial policy sound like antitrust.

[Ben DiFrancesco @BenDiFrancesco]

@auryn_macmillan 🎯

[Ben DiFrancesco @BenDiFrancesco]

If the existence of open weight models kills closed weight labs, it means there's no differentiation between models. A pure commodity. If that's true, the price of tokens is going to be the marginal cost of serving them, due to competition, whether we have open weights or not.

[Ben DiFrancesco @BenDiFrancesco]

@hosseeb @deanwball If the existence of open weight models kills closed weight labs, it means there's no differentiation between models. Pure commodity. If that's true, the price of tokens is going to the marginal cost of serving them whether we have open weights are not (due to competition).

[Ben DiFrancesco @BenDiFrancesco]

@cremieuxrecueil BAN

[Ben DiFrancesco @BenDiFrancesco]

Every now and then I still think about the old times (6 months ago).

[Ben DiFrancesco @BenDiFrancesco]

hell yeah, free cigs

[New York Post @nypost]

Breathing NYC's smoky air is equal to puffing 10 whole cigarettes, expert warns https://t.co/7znoARIY8o

[Ben DiFrancesco @BenDiFrancesco]

@OVTweetmarck