TLDR; someone used some fake graphs and probabilities pulled out of thin air to justify an unwarranted conclusion based on a false analogy to a situation that is far from clear.
So did Strata Vs Llama.cpp, yet Strata pulls 2.5 to 3x the performance of the Llama.cpp codebase it was based on.
If they've improved the architecture further by sharing experts, then this might be plausible.
I haven't even thought about this for more than a few minutes, but in theory your harness could be automatically searching conversation history and project files for items relevant to the current conversation and surfacing relevant things to the agent.
Or the agent could just do that itself proactively.
So many different ways of approaching this problem space.
“The vast majority of this year's net loss, $72 billion, resulted from accounting charges tied to changes in the value of convertible notes and other instruments the company issued to raise funding, the Journal report said.”
Interesting, would love to know the details of this.
Basically the $72B are updating the value of what their convertible notes would be worth if they IPOed today at their current valuation. It's a non cash expense but it does matter because it dilutes all current shareholders and thus lowers earnings per share.
What we need to see is the statement of cash flows.
Whats happening to that cash balance?
People are missing the point IMO - the reinvestment needs is what matters. A) how large are they? B) How will they change over-time? C) How does a slow down in capabilities and closing of the gap by China + price-competition affect their cash flows from operations net of the reinvestment (e.g. will it draw down their cash balance).
You can only hold the perception of increasing capabilities for so long - humans are very adaptive as we are seeing and scare-tactics do not work forever.
I love that you made this, and that you told your story about why. Clearly a lot of thought went into it, both leading up to and during its construction. Thank you!
Like others, I'm skeptical about the business model. That's not really a knock on you (and I wish you the best of luck), but rather with the whole ecosystem of these kinds of apps. There are three big problems:
- Everyone and their aunt is trying to sell me monthly subscriptions for a few bucks. These add up, both in terms of $ and cognitive overhead (yet another subscription I need to remember to cancel if I stop using it...)
- I don't own this system, I'm just renting it. This is a trade-off I'm less and less willing to make for a whole host of reasons including my ability to vibe code my own thing, open source, enshitification (or, perhaps more likely, you deciding to abandon this project in the future) etc
- All the usual concerns about data security and privacy are heightened right now in this era of super-capable hacking agents and super-creepy AI companies
As a user of things, vibecoding is the most exciting thing about AI. I can make anything I want to use, built exactly as I please. As a maker of things, vibecoding is the most damning thing about AI. Anyone can re-make anything I've made, built exactly as they please.
By the way, I've built my own system with AI to manage my todos. So even though I probably won't use yours in the sense of paying for your service, you have some ideas and themes in here that are really thoughtful and I could see working into my flow.
Not only are the models and compute unavailable to the public, they are unavailable to anyone besides the labs, including the government. We better hope the furthermost tip of the jagged frontier stays in purely theoretical mathematics, because if it strays into the physical world then we may find ourselves in a situation where the power of the labs rivals the governments supposedly overseeing them.
reply