That happened at a company I was at 8 years ago. It acquired a company also owned by the major investor. Layoffs started with a month. They whole thing shut down within 6 months.
AI is useful to people who read and understand the answers and who would have eventually come up with a similar result on their own.
They have judgement. They can improve what was generated. They can fix a result when it falls short of the objective.
And they know when to give up on trying to get AI to understand. When rephrasing won't improve next word prediction. Which happens when the situation is complex.
> AI is useful to people who read and understand the answers and who would have eventually come up with a similar result on their own.
I am such a one, and AI isn't useful to me. The answers it gives me are routinely so bad, I can just answer my own questions with a search engine or product documentation faster than I can get the AI to give me something. Often enough I can never get the AI to give me something useful. The current products are shockingly bad relative to the level of hype being thrown about.
> Often enough I can never get the AI to give me something useful. The current products are shockingly bad relative to the level of hype being thrown about.
Yeah, I agree. This has been a big source of imposter syndrome for me lately, since all of this AI coding stuff has skyrocketed
People making wild claims about building these incredible things overnight, but meanwhile I can't get anything useful out of them at all
Something isn't adding up. Either I'm not a very good programmer, or others are lying about how much the AI is doing for them
It also happens when you ask it to do simple things like create comments. At least 400 words and still it will regurgitate/synthesize, often with a "!", the content you're asking it to comment on.
Broadcast indecency rules (like those for a wardrobe malfunction) are a narrow exception and don’t authorize the government to punish political viewpoints; even on publicly licensed spectrum, officials can’t wield licensing power as a cudgel against disfavored speech, because the 1st Amendment forbids it.
No, it wouldn't. The idea of spectrum allocation depends on government intervention. In a free market, transmitting radio waves would be a free for all.
^ no idea why this was downvoted. I'm sympathetic to the idea that most real estate boundaries, even though currently enforced by government, are Schelling points that could mostly persist with private enforcement. But that seems like a tall argument for RF frequencies. Sure, if you were right next to the equivalent of an AM/FM station broadcast antenna, you wouldn't want to bother with trying to reuse that frequency. But if you were hundreds of miles away (where its signal is quite low), you could easily reuse the frequency with relatively small power transmitter.
So the current exclusive use of radio frequencies is very much an artifact of government intervention.
My argument is that generally everyone has access to all the logs. If you restrict the access and add guardrails around it, you can minimize the surface area and also ways it can be leaked out.
If you take a defensive approach towards, you have to assume that some secret is getting logged somewhere. The goal then becomes a way to reduce the surface area or blast radius of this possible leakage.
Limiting access helps, but if you are storing the logs on a 3rd party (e.g. DataDog, CloudWatch), you will still need to assume it can leak through that 3rd party and start rotating.
You're thinking of income tax. Wealth tax is a tax on, well, wealth, rather than just on what you earn. I don't think any US jurisdiction currently has a wealth tax.
Actually I think Zohrans policy is the same as the one enacted in Massachussetss in 2022, the 'millionaires tax'. Its provided 2 billion to the state and theres still more millionaires in the state than there was when it was implemented. I hope we do this in NYC!
