Attracting talent with huge sums of money just gets you people who optimize for money, and it's usually never a good long-term decision. I think it's what led to Google's downturn.
Re:Claude Opus 4.7
by HighGoldstein on Thursday April 16, 2026 @02:47PM
(#47797755)
> I think it's what led to Google's downturn.
What downturn is that exactly?
It used to be THE mecca that all engineers wanted to go to, and they almost never laid anybody off for a long stretch of time. Things changed around COVID. It's very obvious that a lot of their services have gotten much worse- my Google home can barely even set timers anymore and it used to be nearly magical at knowing what song I wanted to listen to even if I didn't know the artist or song name.
They have made great strides recently in a lot of areas, but the massive seemingly random layoffs of senior engineers is very apparent.
Re:Claude Opus 4.7
by __turbobrew__ on Thursday April 16, 2026 @10:19PM
(#47801849)
Google is doing great still. One of the few FAANG I am bullish on over the long timescale.
They certainly could lose 90% of what they were and still be much better than most companies.
> OpenAI has a massive advantage in human capital over Anthropic.
but if your leader is a dipshit, then its a waste.
Look You can't just throw money at the problem, you need people who are able to make the right decisions are the right time. That that requires leadership. Part of the reason why facebook fucked up VR/AR is that they have a leader who only cares about features/metrics, not user experience.
Part of the reason why twitter always lost money is because they had loads of teams all running in different directions, because Dorsey is utterly incapable of making a firm decision.
Its not money and talent, its execution.
Re:Claude Opus 4.7
by staticman2 on Thursday April 16, 2026 @05:23PM
(#47799666)
Are those "smart people you know" machine learning researchers?
Re:Claude Opus 4.7
by __turbobrew__ on Thursday April 16, 2026 @10:18PM
(#47801843)
No, infrastructure engineers. The one who scale the system up so you don’t have to rate limit.
As long as OpenAI can sustain compute and paying SWE $1million/year they will end up with the better product.