I’m locked into what my company allows but that caveman plugin sounds so nice. I usually start my prompts with “hey I’m autistic I want you to be direct with me. When I ask a question that isn’t a secret code to make you do something I’m literally asking a question. Stop writing giant books at me to ask me a simple yes or no question. Don’t explain every concept in my own code base to me, I’m not a layman.” Or some familiar rant lol.
You could always add something like “Be as terse as possible with your responses, using the least amount of tokens possible” in your global rules!
If only it listened to my global rules ![]()
![]()
The newer product of the company I work for is primarily coded agentically. Funny enough, it’s in use at Kaiser Permanente.
I do think there’s a notable challenge in untangling large in use applications to get it easy to use agentic tools with for updates but when it’s been done that way primarily from the start it seems more doable.
Agents are my new development frontier. Anyone writing agents?
Is this spending on AI sustainable?
After hearing how many tokens our company is burning through per month; and how much money that is costing us, I am starting to wonder if all of this can be paid for over time?
Companies like amazon can afford to pay their employees to exclusively use AI; but the other companies are starting to see some serious enterprise expense with little metrics to show ROI. How much does it cost to use Claude to do simple tasks vs. just having the engineer do it?
I read a report from Bain that said AI would need to generate $6 Trillion in revenue by 2031 to justify the data-center price tags. How can business afford even a fraction of that? If business sheds its employees in favor of AI; what will happen to prices? Capitalism says that once the business is locked-in then the vendor will be able to raise prices at will.
I am starting to wonder if AI will be too expensive to replace people? Sort of like a Waymo vehicle. Yes, someday it may work but if it costs $250,000 to make, is it cheaper to operate when compared with a person?
IMO it looks like a giant Ponzi scheme. It’s a handful of businesses that are just paying each other in a circle. AWS pays Nvidia for chips. Nvidia pays AWS for cloud. Claude pays Nvidia and AWS. AWS and Nvidia pay Claude. It’s a tornado of funny money that drives up stock prices. The tornado will continue to spin as long as it has more cows to pick up.
This is the bubble… i have nothing but concern about AI. Yes AI will stay and yes it will outlive the crash but there will be pain inbetween
Consumer inference is not making any money. I believe the reason prices are as ‘affordable’ as they are is because these companies are fighting for market share despite operating at tremendous losses. A flat rate subscription costs a fraction of the raw compute power used (on average). So, it would definitely make sense for prices to increase up at some point. Maybe even prohibitively so.
On a random/different note for this thread..
Unless a treaty is made, I can’t really imagine one superpower slowing down AI development while allowing an opposing superpower to continue unabated. It’s a tough pill to swallow. I think we’re all just strapped in for a wild ride.
In my perpetual state of existential crisis, I sometimes hope that after I come to term it will be like taking off VR goggles so that I can rest assured that none of it was real.
I recently read If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All
I agree with this premise. It’s not even that AI is evil it’s that we feed it incentives in training and have no way to know how they are interpreted and eventually a solve for a request will be detrimental to all of us because the black box we built is simply trying to do what we asked in a way that fits the positive and negative reinforcements we built for it.
In the mean time I guess I’ll do what I’m asked and have my AI agent build my switch configs and review the output like a good AI prompt network engineer… this timeline is dystopian in the least fun way
We need to find a way to send AI to Sunday school.
If ai tries to give me a religious lecture im moving out into the woods off the grid and probably starving to death.
I meant the morals and ethics training that we get regardless of denomination.
You and I may have received very different morals and ethics training in Sunday school ![]()
The problem isn’t AI will act immoral or unethical it’s that we will give it a problem to solve and the solution AI comes up with will inadvertently kill all plant life on earth and ooops no more humans as an outcome. Or will develop is super virus to eliminate a screw worm infestation issue but oops it also eliminates most insects on earth including most pollinators. Same end result we all die as a byproduct of a solution to an unrelated problem statement.
Technically we all are doing this to ourselves but slower just the assumption is that AI can escalate and fast track the destruction we do 100x
i fear the people behind AI don’t really know what they are doing, the programming is inherently flawed because people are writing it and then letting/pushing it to write its own code, and that humanity has a lot of evil people out there that want to use ai to do harm, and the likely hood of anyone stopping it is incredibly low. the world is not 99 percent white hats. there are no guardrails. yes it has potential for good but all i can see is the weapon, curing cancer is at the bottom of the list of things it will do immediately. humanity didn’t use nuclear science to provide safe cheap energy to the world first. it created a bomb with it first.
Woke up this morning to find out 1/3 of my team was laid off this morning. I escaped again. For now.
BBBut they said AI would create jobs… ![]()
Yep that’s right. Now I’m doing at least 2 jobs. Maybe 3. ![]()
This is why I told the recruiters a few years back that I would not work for amazon under any circumstances. They could not understand why I did not want to take a 2 hour ‘skills assessment’ and a personality test so I could work for a company that openly states it will lay off 15% of employees each year.










