On the other hand, much of the code I write is in an industry where training and operations manuals are closely guarded corporate secrets that make up the recipe or soul of a company. The job of the SWE is to deeply understand the processes and procedures that employees follow, and to write code that helps facilitate those and then gets out of the way. A lot of it comes from walking around and seeing how people are actually using the software and what works, and what's a pain point. I've always maintained that the value is in the operations manuals, and the code is just a logical extension of that. But that's where SaaS usually is insufficient because regardless how versatile and broad it is, it doesn't usually encapsulate enough domain knowledge, let alone the proprietary stuff.

Since I started using AI tools to assist, I've found a lot of both utility and frustration revolves around my use of these nouns in prompts (in the context of, e.g. "during the 'quickpay' confirmation phase..."). When the bot settles into understanding these nouns, it seems to get a better handle on the architecture as a whole. When it suddenly forgets them and has to go figure out what they mean by scanning the code base, I know it's about to do something staggeringly redundant and stupid.

Which is also to say it's a cheap bet that anyone with no reputation can afford. Hence, not believing doomsayers mean what they say is a sort of societal hedge against people flooding the zone with doomsday scenarios about everything.

The problem with this inversion of your first statement (that violence is not the answer), which everyone justifying violence in this thread seems to forget, is that there is always someone who feels this way about anything.

The words and narratives of Martin Luther King, Jr., for example, caused so much fear and uncertainty and anger in some people that they thought their only option was to commit a horrific crime.

Someone responded to you below saying if you feel that peaceful revolution is impossible, then violent revolution is necessary. That person feels that they are on the side of justice. What they forget is that so does everyone else.

The reason revolutions rarely stop where a reasonable person would want them to stop, and instead continue into eating their own and counter-revolutions, is that once you say that it's understandable to take out a proponent of (X narrative), there's no end to the number of people who will justify violence in the same way against any other narrative as well.

We can all well think that Altman is opening Pandora's Box, but that doesn't justify opening it ourselves, or giving a pass to wannabe revolutionaries who would.

In retrospect, too, we can say that the assassination of Hitler had it succeeded would have been a good thing. We can say that the elimination of the ayatollah by the US was a good thing. What we cannot say is that an individual's perception gives them a right to commmit murder.

https://news.ycombinator.com/item?id=47392550

Today I checked my spend again and found that my limit had been raised to $200,000 per month. I'm not complaining, but that seems a little strange to me. My total all-time spend is about $400 right now. I don't think they've checked my credit score. Not to sound paranoid, but ... okay, I'm going to sound paranoid for a second. Is Claude reporting back that my sports prediction software actually works? Who the ** gets a $199,000 increase in their monthly limit when they pass $400 in spend?

To put it another way, they just gave me a limit that surpasses my annual net income. Or is that just the normal jump, from $1k to $200k that's baked into their billing system based on some frequency of usage metric?

Comments URL: https://news.ycombinator.com/item?id=47597238

Points: 2

# Comments: 4

Can you feel when your agent has just compressed or lost context? Can you tell by how it bullshits you that it knows where it is, while it's trying to grasp what was going on? What's your emotional response to that?

Comments URL: https://news.ycombinator.com/item?id=47516319

Points: 4

# Comments: 16

Previous: https://news.ycombinator.com/item?id=46781990

Comments URL: https://news.ycombinator.com/item?id=47508312

Points: 1

# Comments: 1