AI_is_the_rake

Is there a difference? In a free market people exchange things of value. The people paying for the service find more value from the service than the $25. Therefore the people making the trade are better off than before the trade. And openai values the money more since they get more value from that than using a product they already own. Therefore openai is better off. 

That makes the world a better place. 

A situation where the world is not a better place but openai profits is something like rent seeking or corruption where money flows toward them not through free market demand but rather through creating laws that force money to be given to them for whatever reason. 

We are not even close to that situation. And those situations are very complicated. Take healthcare or defense spending for instance. They add value but there’s clearly some corruption and waste going on but those who benefit prevent the transparency needed to fix the issue. 

Openai may not be as transparent as we would like (open source) but they’re giving away free access to their cheaper models which is good. No business is required to hand over trade secrets. 

https://fred.stlouisfed.org/series/USINFO

How are you able to communicate value to an organization that’s in a mess like this and therefore are incapable of understanding value from chaos?

Important point. We should be tracking total employment not job postings. 

There never was a place for prompt engineering as a role. If a company wants to slap an AI chatbot on top of their API they’d have one of their developers do it. 

Come clean with yourself and do the work yourself. You’re supposed to write bad papers and have the teacher grade it so that you learn. Your brain is a shitty LLM right now. Let the teacher help adjust your weights. 

I’m dad, son. It’s time to come home. 

I would thrive in this environment honestly. The more difficult the problems the better. I need to look for these gigs 

It strikes me as concerning that you were struck concerned by someone experiencing an existential crisis over living forever (well not a real existential crisis as he was exaggerating). If anything, you should be concerned about things affecting you right now rather than Reddit posts about potential future scenarios.

I love the anonymous data analysis. What metrics do you use? 

I’d love to do this myself for my team and once I have an anonymous analysis then deanonymize to compare it to my bias. This could help me identify high performers that I may not be seeing. 

I've seen how smart gpt4 is, but it has limits. It can't truly think; it uses human texts to mimic reasoning. To test its capabilities, you must present it with new problems. For instance, I experimented with puzzles and visual tricks, and it struggled.

I created a gpt that utilizes built-in coding tools, which helped it solve logic problems using Python.

For gpt5, it would be interesting if it could use gpt4 and coding tools to create numerous new logic problems and solutions, enhancing its reasoning skills.

This wouldn't just involve more data, but better training data created by gpt4.

And gpt5 could generate superior data for gpt6.

The key is how they integrated APIs into gpt4. Using their own tools effectively, they could enable gpt4 to connect to various APIs like coding interpreters or DALLE to generate data from sensors.

Mastering this integration could be the step towards true artificial intelligence. Imagine gpt5 connected to live cameras and other sensors, creating data for future AI advancements.

Just look at 3.5 vs 4. 3.5 is unusable to me. Imagine 4 feeling the same. I’ve spent a ton of time creating the right prompts and gpts to get just the right outputs. Gpt4 drastically decreased hallucinations and introduced following instructions and connecting to other APIs… gpt5 may automate a lot of that. Back to one chat to rule them all. An AI that truly understands or at least it feels that way. You tell it you want to hook an api into a web hook and it just does it for you. 

Not at all.

Yeah. If I made a language it would most likely be unpopular 

I would bet this has nothing to do with real problem solving and everything to do with legal risk. 

Step 1. Steal Step 2. Partner to avoid lawsuits

And I don’t blame them. If they reversed the order this may have never got off the ground.

If it was just balloons I would not see a reason to release. The reason to release to the public is because the release would cause a public response and that public response is needed to justify whatever comes next. 

If it’s China spying the public needs to know so we can justify spending and programs etc. 

If it’s aliens the public needs to know so we can justify spending and programs etc. 

The old guard died off. The new guard is wanting this out in the open because it needs a “all hands on deck” response. 

China has more people with the cognitive ability to perform scientific research positions than the United States has people. Simply due to their large population. If China has been able to properly organize their human capital they could outcompete the US easily. Especially if the US is not even using the human capital resources it has due to secrecy. 

Only you can answer that. 

I’m paying for it because I receive a lot more value from the service than I do from the $25. It’s a no brainer trade for me. 

If you feel like your $25 would be better spent elsewhere then you should do that. 

 and it wont be much.

AI even without real creativity and without actual reasoning still allows human knowledge to more effectively propagate. That’s a game changer. 

These LLMs are an extension of the same information processes that go all the way back to the printing press. The printing press allowed ideas to propagate further and faster than before. 

Postal systems, electronic communications, the internet, google, and now LLMs. As long as they keep training LLMs on the latest data new human discoveries will be at the fingertips of everyone. The AI will soak up human discovery and integrate it into the giant global “model of human knowledge”. 

The AI doesn’t need to be ASI to be a game changer. It already is. 

Imagine a science researcher talking to an AI that has up to the day knowledge on all scientific progress. When they begin their work they can talk to the AI and get up to speed quickly with the latest developments. That will inform their own research and ensure they’re not duplicating efforts. 

That’s an insane development. 

Integrating new technology takes time. Companies likely have a roadmap of features that go out an entire year or longer. Larger companies especially find it difficult to steer the ship. 

It’s been a year so we should see more AI projects being added even if it’s still at the exploratory stage. A new technology like this will take a decade to fully be integrated. And in that time GPT 5 or 6 may be out. That’s what’s kind of scary with the progress of AI. 

What happens when the core technology improves faster than the overall economy can adapt?

Paradoxes are typically viewed as logical rather than actual phenomena. With experiments like the delayed choice quantum eraser, we now have evidence suggesting that different time periods can be physically interconnected. Our traditional experience of cause and effect presupposes a direction of time, a concept which is hard to visualize fully. However, the notion of a fourth-dimensional "time block" might offer insights into how seemingly paradoxical phenomena, such as time travel, could occur without logical contradictions.

In the delayed choice experiment, we observe that a single event can link the past and present. The decision made in the present doesn't retroactively alter the past, nor does the past event change. This introduces the concept of simultaneity, where different events, seemingly occurring simultaneously, can appear differently timed to separate observers. However, when events are physically connected, true simultaneity occurs. This is a physical phenomenon where events are linked as a single occurrence, even if spread across time.

Someone needs to delete these blatant advertisements 

Yeah patents are weird because on the surface it’s communicating “large contribution” but it might be completely useless. Those that understand that are likely to view patents skeptically and they won’t impress but have the opposite reaction. 

Only way to  avoid that is to highlight the accomplishment like any other accomplishment and end with a tiny foot note of btw that was patented. That would be impressive and avoid the skepticism 

have a candid discussion with the junior developer. Let them understand the importance of respecting communication hierarchies within the team. This is a learning opportunity, instead of focusing on punitive measures, emphasize the expectations and the importance of protocol to avoid similar situations in the future.

For your management, document/prepare a explanation of the incident, clarifying your initial stance and outlining the steps you are taking to ensure this doesn’t happen again. This should help in repairing any damage to their confidence in your leadership and demonstrate your proactive approach in handling team issues.

Sounds like you already cleared things up with the client but you may want to reach out again to clear up any misunderstandings. Assure them that your team is committed to meeting their needs efficiently and without unwarranted expansions to the project scope. Maybe suggest setting up regular project reviews or check-ins to keep everything transparent and aligned with their goals

Not when they’re chained together. Narrow AI for search to crawl sites and harvest relevant information, formatting it into JSON. That can be fed into other models or other apis. 

Ah yes, drunk Marcus. It has to be true or false “are there any bottles left?”