I'd say you only pay 10% of the cost when you finish/launch the re-written software. The other 90% of cost comes later in the format of operations, deprecation and migration, people both users and devs, community friction and learning cost.
I think in time people will realize harness is essentially a more complicated .vimrc or .zshrc;
And yes, you can install gigantic plugins in those places - e.g. Codex; but the point is everyone will have exactly what they have customized towards. The more atomic a building block is, the easier it can be adapted into any kind of configuration.
I think the pain of selling a harness is if your target market understand what a harness is, then they can build it to exactly how they'd like it without much effort. If they don't, then the harness wouldn't be very useful to them in the first place.
This looks like colloquial mandarin translated to english word for word.
This is almost exactly how mandarin is spoken in colloquial communication, context is almost always omitted if provided once, and grammar/tense is mostly non-existent either. Check out Singlish for something similar but understandable if you are not a mandarin speaker.
I'd much prefer this over agents that enthusiastically implements whatever they are asked to do and make up whatever information they think is missing.
One of the biggest reason I migrated to Codex. I'd like to see each agentic CLI process as a simple building block, and having 100 claude session means I sometimes run out of memory (and I have 256GB of them). Codex never have this problem.
The communication is more direct too with GPT5.6 family of LLM. No more belt and suspenders.
It seems like they tried to remove guidance from multi-agent system. And I think it's going to fare as well as removal of guidance from single-agent interactions.
In my experience, no matter how many agent runs for a single goal, one of the pre-requisite is clear and concise communication so that LLM are left with as little freedom in the matter of arbitrary choices, or "taste". When they are given too much choices in this regard, the outcome almost invariably bad.
I think this has to do with LLM lacking in purpose - a dictionary and encyclopedia can have all the worlds knowledge but it is completely neutral. A reflection of your commands from an LLM is similar to a lookup process despite it can be made to "do things". This purpose is likely not something that can be given to the LLM in the current format.
It feels like a great product but targeted at a surprisingly transient customer base.
The cycling guys won't buy it, they want the watts from their legs.
The commuters will just get an ebike eventually.
The getting started crowd will eventually become one of these 2 or stop riding a bike altogether. Since I didn't see how much this cost, if it cost as much as a commuter bike, I really don't know who would buy something like this.
There are not enough human 25 years old in history for this many 9s. You had 21 nines, which would only make sense if there were 10 billion billion humans ever and that he is the only one. However, the current estimate is there were 107 billion humans ever alive in history, so your statement is impossible even if the estimate is off by 10000 fold.
Your failure here is that you are using normal math, not Effective Altriusm math. That number should not only include past/current 25 year olds, but also all possible future 25 year olds, including all 25 year olds simulatable by a future powerful AI. In that context, the estimation is quite modest.
It's a rhetorical device that I have yet to see used to good effect. The kind of rhetorical device that undercuts the argument rather than strengthens it.
You're so caught up in irrelevant details like a rhetorical device that you've missed that this guy's fund is up 80% for the year and his investors are having a GREAT year. Worth reflecting on.
And that 90% could be an underestimate here.
reply