This is the hype train, claiming they can and will do so much more than be junior software engineers.
Why would they want a slowdown while simultaneously getting ready for an IPO?
Because for one, these things can’t and won’t do so much more than be junior software engineers because a complete lack of understanding of how the human mind does not inhabit the body or pilot the body, but the mind separated from the body loses wholeness. That’s going to blow up these IPOs if they can’t actually deliver, so the slowdown is the industry all shaking hands and saying “we need more time for this charade to hold up” when actually none of them will slow down, they’ll just be trying to hide how little forward progress there is to be made left with these kind of models and without moving to something even more complex. They need the con to last as long as possible in hopes of delivering something useful at all to justify the massive increase in surveillance capitalism that citizens obviously do not want. Their IPO will crash if they get found out to be frauds right after going public, and they aren’t satisfied enough with a quick pump and dump, they all want to be the next Musk. So they need a global slowdown to justify how things have actually hit a wall in terms of making LLMs useful to every day people.
Anthropic’s opus 4.8 is well beyond a junior dev. Most of the time it outperforms senior devs. Then there’s mythos we don’t have access to yet. We’re fucked. Even if human R&D stops now
This is the hype train, claiming they can and will do so much more than be junior software engineers.
Why would they want a slowdown while simultaneously getting ready for an IPO?
Because for one, these things can’t and won’t do so much more than be junior software engineers because a complete lack of understanding of how the human mind does not inhabit the body or pilot the body, but the mind separated from the body loses wholeness. That’s going to blow up these IPOs if they can’t actually deliver, so the slowdown is the industry all shaking hands and saying “we need more time for this charade to hold up” when actually none of them will slow down, they’ll just be trying to hide how little forward progress there is to be made left with these kind of models and without moving to something even more complex. They need the con to last as long as possible in hopes of delivering something useful at all to justify the massive increase in surveillance capitalism that citizens obviously do not want. Their IPO will crash if they get found out to be frauds right after going public, and they aren’t satisfied enough with a quick pump and dump, they all want to be the next Musk. So they need a global slowdown to justify how things have actually hit a wall in terms of making LLMs useful to every day people.
Anthropic’s opus 4.8 is well beyond a junior dev. Most of the time it outperforms senior devs. Then there’s mythos we don’t have access to yet. We’re fucked. Even if human R&D stops now
It’s both smarter in many ways, and extremely dumb in others, even in a really good and secure tool harness (which is rare).
It’s a great assistant and task finisher, but the architecture is just fundamentally not suitable for handing any significant responsibility.
IBM 1979 training manual
I disagree. We should replace the C Suite Immediately. Lazy, entitled and unaccountable and they are paid WAY beyond their value.
Even worse: nested API calls to ever-changing black boxes with literally randomized outputs.