Coley 🌊

Forecasting and Trends. Data led, rarely wrong.

Ex Chartered Investment Manager turned Business Intelligence Leader.

[Coley 🌊 @WeAllFlow]

@Liv_Boeree You might have a point but please don't use GBeebies propaganda as a source. 🙏

[Coley 🌊 @WeAllFlow]

@yongfook RLS. People who access the API can only access data that relates to them. You have data for everyone and everything.

[Coley 🌊 @WeAllFlow]

Most people on here on in a bubble.

In the corporate world, AI has only made employees lives easier.

There's rarely an increase in revenue or decrease in costs.

Most big companies can't quantity the benefits of AI.

If you can help them do this you can print money.

[Coley 🌊 @WeAllFlow]

@BabaDev_ Post on twitter.

[Coley 🌊 @WeAllFlow]

@Bencera Approximately $34 billion was an accounting charge associated with financing that could convert into shares, rather than operating cash expenditure.

Nearly a quarter of revenue reportedly came from two customers though which is a bigger risk.

[Coley 🌊 @WeAllFlow]

@ABridgen He's 100% my new favourite word, a Quockerwodger.

[Coley 🌊 @WeAllFlow]

Most 'builders' are building tools for other 'builders' because none of them have any domain knowledge in anything else.

99% are making agents to replace tasks they weren't even doing in the first place.

[Aanya @xoaanya]

Serious question:

are the people making money off AI actually building? or just selling courses about building

[Coley 🌊 @WeAllFlow]

@BasilTheGreat Him and his mates made this country not worth fighting for. 🤷‍♂️

[Coley 🌊 @WeAllFlow]

@EbiThinks AJ

[Coley 🌊 @WeAllFlow]

@Liv_Boeree I'm getting pushed loads of accounts posting false stuff about the UK.

Definitely another agenda on the loose that's worked out the algorithm.

[Gossip Goblin @Gossip_Goblin]

Victory By Any Means

[Coley 🌊 @WeAllFlow]

@real_shirelass Possibly one of the worst takes I've seen. 😂

[Coley 🌊 @WeAllFlow]

@NextGenAi5 As a data analyst, I approve of this list.

[Coley 🌊 @WeAllFlow]

If you remove training frontier models, Anthropic are actually very profitable.

If only there was some sort of global call to slow down frontier AI model training in case of an IPO or something. 🤔

I've reconstructed their costs based on leaked Q2 ratios on $11.5B revenue:

Lab
🔹39% - Training compute

Product
🔹21% - Other cash OpEx
🔹17% - Inference compute
🔹10% - Partner revenue share
🔹8% - People / cash comp
🔹5% - Adj. operating profit

NB. These are reconstructed from leaked ratios on $11.5B revenue, not official financials.

[Coley 🌊 @WeAllFlow]

Dario's track record:

Aug 2023: A model that talks like “a generally well-educated human” in 2–3 years.

Oct 2024: “Powerful AI”, what he calls a “country of geniuses in a datacenter” “could come as early as 2026, though there are also ways it could take much longer.”

Jan 2025: Systems “better than almost all humans at almost all tasks” in the next 2–3 years.

Mar 2025: “AI is writing 90% of the code” in 3–6 months. In 12 months, “AI is writing essentially all of the code.”

He later said it was true “on many teams,” and that “70, 80, 90% of the code written at Anthropic is written by Claude.”

May 2025: About half of entry-level white-collar jobs gone, unemployment 10–20%, in 1–5 years.

“This is about to happen.”

Later he said: if you automate 90% of a job, people do the remaining 10%, and that 10% expands.

2026: “Country of geniuses in a data center” still a 1–3 year guess; high confidence it happens within 10 years.

What do you think? Optimistic or delusional?

[Dario Amodei @DarioAmodei]

We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.

Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.

You can read the full post here: https://t.co/OGyPb7yaYt

[Coley 🌊 @WeAllFlow]

For the right amount of money because he knows the IPO is cooked.

[Coley 🌊 @WeAllFlow]

@FreshFifth How else do you want me to say it numb nuts?

[Coley 🌊 @WeAllFlow]

I GUARANTEE this is all about restricting access to local models.

This is about holding onto power, not safety. These guys couldn't care less about humanity.

Remember this post when the U.S. introduces an "AI Safety Bill" later this year.

[Polymarket @Polymarket]

BREAKING: Sam Altman backs Dario Amodei’s call to slow frontier AI development.

[Coley 🌊 @WeAllFlow]

@opropriofaka 100% in the cost case, they're getting their excuses in early.

With banning local models that's one of things they're desperately trying to protect with this narrative.

[Coley 🌊 @WeAllFlow]

👇

[Coley 🌊 @WeAllFlow]

Btw if I get to 500 followers by the end of Monday I'll make a video showing why and what will happen next.

[Coley 🌊 @WeAllFlow]

The only other explanation is cost.

These companies are all hemorrhage money and the models are progressing faster than infrastructure.

Pretending to slow down frontier models is a collective agreement to curb costs until compute becomes cheap enough. Again, not safety.

[Coley 🌊 @WeAllFlow]

Btw if I get to 500 followers by the end of Monday I'll make a video showing why and what will happen next.

[Coley 🌊 @WeAllFlow]

I GUARANTEE this is all about restricting access to local models.

This is about holding onto power, not safety. These guys couldn't care less about humanity.

Remember this post when the U.S. introduces an "AI Safety Bill" later this year.