For human in the loop to be effective, the human needs to actually be performing some substantive action, giving real guidance and critique and pushback. If the human only ever accepts the default plans then not only is there no understanding but the agent should learn to stop asking. It is not learning anything from the human, after all.
One thing that I look at is pushback rate: what percentage of the agent's proposals are rejected or critiqued? If it's below 5% I have found I have gotten too credulous and I am no longer closely following. Danger! If it's above 50%, I have clearly not given the agents sufficient context to perform the task and need to update my harness and instructions.
Who watches the watchers? I can imagine a guard dog process that halts the session to yell at the human if it detects complacency: if the human is providing too few tokens per minute of new context relevant to the task.
That does seem to be the path Apple is following here. Have a local model that can answer most things and then have a fallback of cloud options when they request is too complex. The cleverness of this strategy has been overshadowed by the incredibly poor quality of their local models. It will be extremely interesting to see what next month holds and whether Google helped fine tune an Apple specific Gemini / Gemma model for their devices. Bonus points, of course, if they unveil the M5 Ultra Studio with half a terabyte of RAM to be a local "cloud model" (the true fantasy here of course would be Apple building something a little like openclaw where from your phone you could give commands to your Home Apple server). They could probably get away with charging $20k for it if it has sufficient tok/sec. If that happens and is successful one could imagine a straight line path in the next two generations to bringing the cost and form factor down to the point where some of the form factor of an Apple TV becomes everybody's home inference server / agentic HQ. Sovereign AI for everyone!
Reminds me of this ICE contract for billion dollar detention facilities … to some random unknown company registered to a single family home with no working phone number or website
It's nothing new, the Pentagon has failed 8 audits in a row. In just one year (2022), $220 Billion dollars of their spending were totally unaccounted for.
that audit has nothing to do with the president of the united states enriching themselves and frieds. If you want to go look at that audit, the summary usually is: the DoD can't place a bunch of black money going to CIA programs to do whatever shit they get up to.
And by "places orders" we mean "helps TSMC acquire plots of land on which their next facilities will be constructed" kind of level of scope, timing, and commitment.
In my experience, the corporate-speak "partnering with" can mean almost anything.
Apple gives TSMC a billion dollars to build a cutting edge fab dedicated to making Apple's chips, a deal they repeat several times over more than a decade? Partnership.
Youtuber takes $300 to read an ad, giving viewers a 10% discount code? Also a partnership.
There's partnering with Apple for several decades where they plan years in advance and pay billions without fail, and there's partnering with OpenAI where Sam Altman commits to giving you a Trillion dollars provided you can deliver all that ram up front and he can give you an IOU he got from Oracle who got it from Nvidia who got it from OpenAI. These are different things.
No, but they can and so can Apple if it becomes critical, I don’t know what is more critical to Apple than replacing Intel, Qualcomm, or Nvidia, but memory probably is number four on the list which means it probably is something that will be addressed?
Note that this is how the FAA has been doing its written tests for years: you go to a proctor test facility and through a metal detector. The entire thing is videotaped.
My read on the book was "humans are really good at telling if you genuinely care about them or not and will respond well to that, so you should genuinely care about the people around you, and good things will result from that overall, especially if you're not super mercenary about it."
Bill & Ted said it most pithily: be excellent to each other.
Serious question - what is your definition of "genuinely care"? The Carnegie example doesn't show "geninuely care" to me. It's nice, I think I should do it. Give people random but geninue complements. It's nice. It costs nothing. It makes both of us feel good. But, is that "genuinely care"?
I ask because I'm bad at conversation. I hear this "genuinely care" and I just, usually, can't get myself to do it. I don't care. I would like to have a nice conversation and I try to care in the moment but the odds are pretty high that 5 minutes after it's over I'll not even know their name and move on with my life. That's not "genuine care" to me.
You can genuinely care in one moment and forget about them five minutes later, that's ok. Part of making conversation is also stopping the conversation when you are done, not letting it bleed out. My favorite is: "I really enjoyed this conversation, now I'm going to read my book".
I find it nice to connect to strangers in real life, if only for a moment. It can be about something silly as sharing a very bright bird you see while waiting for the bus. It can be giving two dollars to the woman in front of you at the grocery store, cause she's short.
Also, having this connection with people when it is about nothing (small talk) helps build communication skills you need when it is about something.
I genuinely hope you will get answers on your question, maybe even in this thread. But I'll also forget about it in two minutes.
That's kind of the point. Ask yourself: which people would you genuinely be excited to make a little happier? (through a compliment or otherwise) Whose opinion are you keen to carefully listen to and consider? Who do you like enough such that you will want to put in the effort to remember their name?
I think the idea is that if the stranger on the bus has a haircut you genuinely find to be wonderful: tell them about it. You don't need to force yourself to be nice, just take action on the things you're genuinely excited to do.
And if you don't ever want to be nice to people, then you have some digging and reflection to do (including about if/when you are nice to yourself).
I'd argue that there is a very strong value in doing something good, not just because it's genetically or socially imprinted on you, but because you actually decide to do it.
This applies to everything, there is no merit in being good at something just because you were born that way.
But the question was different: it wasn't "can I get good at this flowery small talk if it doesn't come naturally?", it was "is flowery small talk genuine care?"
I would posit that no, it is not. And it's not even unambiguously a good thing. There are plenty of cultures where people are described as cold until you get to know them, but once you do - they'd die for you. To me, that is genuine care. The American "Hiii! How ARE you? I don't actually care if you keel over and die!" approach feels fake.
Giving someone, as in this thread, a genuine compliment that you mean sincerely isn't "flowery small talk" and it's sort of depressing that you think that it is.
No one in this thread is talking about your example except for you, and it would perhaps do you well to reflect on why you read things that way.
I’m not so hung up on the semantics. The fact that you’ll likely never meet someone again can render an act of kindness towards them, no matter how small, more meaningful, not less.
I think a reasonable proxy for "care" is genuine curiosity about who a person is, how they came to be where they are, what makes them tick, coupled with a general desire for positive things to happen to them in their life.
If you believe that most people have things about them that are fundamentally interesting, you will put effort in to find those things, and you'll generally be successful in finding them. If you have a belief that most people are fundamentally uninteresting, you will not put in the work, and your beliefs will be validated because you will fail discover anything of interest in most people around you.
The act of earnestly wondering about a person is very flattering - this will help people open up to you and share their interestingness. If you can layer that with an additional thought of what useful thing you might be able to do for them, such as another person who might be able to help them, or a piece of advice you can give that might save them time, you will develop a reputation as a person one should talk to.
This then will increase your hit rate on being useful which is a nicely positively compounding cycle.
I think when it comes to small talk and small moments with people, caring is meant literally. You care that they have a decent day, a brief nice moment. So in carnegie's example, he notices that the fellow looks bored, and he sees an opportunity to take care of him, in the form of a compliment.
I think your comment reflects that you're waiting for someone to say or do something which will cause you to care. And that's effectively waiting to get something from them. You need to cultivate the sense that everyone in some sense has the same daily struggle, and be the bigger person who strives to alleviate some of that loneliness and suffering in others.
> Give people random but genuine complements. It's nice. It costs nothing. It makes both of us feel good.
I'm worried it'll come off as creepy/weird so I never do. I've seen the power of other people doing it, but I cannot do it free from that worry so it's always gonna be off vOv
The only path to cool is through cringe. You can do it!
I think a lot of people after a random compliment might be wondering if the other person is trying to get something out of them, like a date or a business deal. When it becomes clear that you're leaving it there with no expectation of reciprocation they can truly internalize the compliment.
Its not really about "genuinely caring" about the individual scenarios or things. Its about genuinely caring about the lives of other people and things which you may not have direct experience, aptitude, or a prexisting joy for.
I genuinely care about my friend. He's really into bee-keeping. I don't care at all about bees. But he cares about it, so I ask questions because I care about him. I have now learned enough about his bee-keeping to be legitimately interested in whether, say, his bees survived the winter or to be upset with him that an invading swarm killed them.
The simple answer to your question, I think, is that you probably can't "make yourself" care about a specific thing at the drop of the hat. But if you care deeply about other things, especially tangential things, it's relatively easy to learn to care about new things you learn about.
Without doubting anything about your intent, I feel like I am the guy who shows interest in other people, but other people show very little interest in me. Frequently, I am listening to someone talk about someone they care about, and I attentively listen, and try to ask friendly, encouraging questions. At some point, my mind begins to think: "Can I just replace myself with a cardboard cut-out and will they keep talking about themselves?" Sadly, I think the answer is yes. I don't know what to do about it.
1. You don't care about X until you do. Like, you can go for years without worrying cholesterol. And then you can have a reason to care about it and all of a sudden you do. The reason can come from something that forces your hand or just because you take an interest in a subject.
2. Altruism. Think less about care and more just doing without expecting anything back. People notice, especially with selfless conversation.
For me, I find most things can be fascinating. There are so many domains I have zero personal, surface-level interest in, but have nuances that are super interesting.
When someone else has that spark, and their eyes sparkle, and they beam as they talk about "their interest"? Idk, I love that. It makes me feel good to hear them. I feel like we both come away better for the conversation.
> How can you make yourself genuinely care about something you don't care about? It sounds as plausible as changing your own sexual orientation.
Most people don't care about the gym but they care about their health and their health as they age so many learn to care about going to the gym even if they don't love every minute of their gym time. I'm one of those people.
Not sure what the downvotes are for on this one. It depends a lot on what "genuine care" is supposed to mean. If you want to interpret that as a subconscious feeling then you're right. Feelings aren't normally controllable and calling them up on demand is pretty much impossible.
That being said, if you go through a bit of game theory and apply it to the real world - the experience of the last few millennia of recorded history is the strategy most likely to get people what they want is lots of communication and setting up win-win deals for everyone. Someone who reliably offers win-win deals has a natural advantage over the more common person who thinks in terms of win-lose deals. Communities that make a habit of setting up win-win deals for their members have an overwhelming advantage over those that don't. If you tap in to that type of thinking it tends to translate into taking a real interest in how other people are going because it is easier to set win-win deals up if you know what their problems and goals are. And a sensible sub-strategy is making sure to be as kind as possible to everyone to get into the habit of thinking empathically and keep channels of communication as open as possible.
So if "genuine care" means you literally feel something... nobody has much use for your feelings, we can't tell what your feelings are anyway and you probably can't call them up on demand. If "genuine care" means you try to figure out what other people want and then help them get it then that's simply good strategy and most people should find their way to it if they think about it for long enough. Some people have to think a bit harder than others and there are a few rare maniacs who really just want to cause pain and suffering. The maniacs are bad news.
The pricing models that are published on AWS' website almost certainly have almost nothing to do with the pricing models that are discussed behind closed doors for a $100 billion commitment.
Of course not, but unless they’re getting the sweet heart deal of a lifetime from Amazon of all places, it’s still a hogwash. We’re talking about enough capital to build their own fab and a dozen datacenters*. This deal isn’t going to be buying existing capacity because that’s already stretched, it will be paying for new buildouts.
Afterwards Amazon will be milking the machines these commitments buy for nearly a decade. That tradeoff makes sense at a small scale (even up to $X00 million or even billions), but at $Y0 or $Z00 billion?
Color me skeptical. There are plenty of other side benefits like upgrading to the newest GPUs every few years, but again we’re talking about paying for new buildouts with upfront commitments anyway.
* obviously the timelines, scientific risk, and opportunity cost make this completely infeasible but that’s the scale we’re talking about. It’s a major industrial project on the scale of the thirty year space shuttle program (~$200 billion).
Having just checked my child in for their doctor appointment, 90% of web software would be dramatically improved by using very boring best practices and readable and accessible web practices.
GPT 5.4 is the surly physics PhD post-doc who slowly and angrily sits in a basement to write brilliant, undocumented, uncommented code that encapsulates a breakthrough algorithm.
Opus 4.6 is the L5 new hire SWE keen to prove their chops and quickly turn out totally reasonable code with putatively defensible reasons for doing it that way (that are sometimes tragically wrong) and then catch an after-work yoga class with you.
> and then catch an after-work yoga class with you.
That's cute, but do you mean something concrete with this, aka are there some non-coding prompting you use it for that you're referring to with that or is it simply a throwaway line about L5 SWEs (at a FAANG).
(FWIW, I find myself using ChatGPT for non-coding prompting for some reason, like random questions like if oil is fungible and not Claude, for some reason.)
One thing that I look at is pushback rate: what percentage of the agent's proposals are rejected or critiqued? If it's below 5% I have found I have gotten too credulous and I am no longer closely following. Danger! If it's above 50%, I have clearly not given the agents sufficient context to perform the task and need to update my harness and instructions.
Who watches the watchers? I can imagine a guard dog process that halts the session to yell at the human if it detects complacency: if the human is providing too few tokens per minute of new context relevant to the task.
reply