I still don't get why some people think just talking to an LLM which generates code without reviewing the code or doing some hands on to tweak it or just generally understanding it means they don't need an IDE.
Especially when 45% or more of the code is 1) insecure (proven), and 2) messy and probably unmaintainable or simply unknown by the "developers" (cognitive debt.)
Meanwhile, idiots are considering "tokens burned" as some sort of "productivity measure" - exactly the same as "lines of code" back in the day which was never valid, except for corporate bean-counters.
What happens when open source developers can't use an IDE because they've all gone corporate? Add to that the "age-verification" nonsense and "the model is too powerful to release", both of which appear to be intended to destroy open source as a platform by regulating it out of existence at the behest of the corporations.
The only thing the Anthropic Mythos brew-ha-ha proves is how BAD so-called "software engineering" actually is. It never was "engineering", it was a cottage craft. The computer security industry is clawing its eyes out (pun intended) over the "OpenClaw" fiasco and agent security in general.
I've long said that software should be designed by actual design engineers following real engineering principles and checked by AI expert systems to catch mistakes in usability, reliability and security. Then a button is pushed and an AI generates machine code for any architecture you want. "Programmers" should be automated out of existence.
But I don't think the current LLM technology is the answer to that. It simply isn't reliable or secure enough to power the entire software needs of the world.
The first part discussing the Cursor deal and Roomote was sharp. But the next part about Kilo wasn't. In my humble opinion, Kilo is not currently in a safe spot. The market is rough yet Kilo's value propositions aren't clear/essential enough to the devs. And the quota (Kilo Pass) is not convincing, at least to me. You guys lack a differentiator to stand out.
Kilo code thinks wider. Kilo code thinks about developers
By the way, $60 billion - or even $10 billion - for an IDE company?
Ridiculous.
The upcoming Iran-war-caused global Depression will put paid to that nonsense.
I still don't get why some people think just talking to an LLM which generates code without reviewing the code or doing some hands on to tweak it or just generally understanding it means they don't need an IDE.
Especially when 45% or more of the code is 1) insecure (proven), and 2) messy and probably unmaintainable or simply unknown by the "developers" (cognitive debt.)
Meanwhile, idiots are considering "tokens burned" as some sort of "productivity measure" - exactly the same as "lines of code" back in the day which was never valid, except for corporate bean-counters.
What happens when open source developers can't use an IDE because they've all gone corporate? Add to that the "age-verification" nonsense and "the model is too powerful to release", both of which appear to be intended to destroy open source as a platform by regulating it out of existence at the behest of the corporations.
The only thing the Anthropic Mythos brew-ha-ha proves is how BAD so-called "software engineering" actually is. It never was "engineering", it was a cottage craft. The computer security industry is clawing its eyes out (pun intended) over the "OpenClaw" fiasco and agent security in general.
I've long said that software should be designed by actual design engineers following real engineering principles and checked by AI expert systems to catch mistakes in usability, reliability and security. Then a button is pushed and an AI generates machine code for any architecture you want. "Programmers" should be automated out of existence.
But I don't think the current LLM technology is the answer to that. It simply isn't reliable or secure enough to power the entire software needs of the world.
The first part discussing the Cursor deal and Roomote was sharp. But the next part about Kilo wasn't. In my humble opinion, Kilo is not currently in a safe spot. The market is rough yet Kilo's value propositions aren't clear/essential enough to the devs. And the quota (Kilo Pass) is not convincing, at least to me. You guys lack a differentiator to stand out.