bmks270 said:

---

AgGrad99 said:

---

Buck Turgidson said:

---

I can see how this would wipe out many entry level engineering jobs....

---

This, to me, is the crux of the issue.

How do we train people for more higher-level/expanded jobs, if the entry-level jobs are taken up by Ai?

It's going to cause some massive growing pains.

---

Only human jobs will be entertainment and services where humans want human interaction, and in frontier technology innovation where AI doesn't have any training data or other examples to pull from and copy.

Engineering that is codified like a lot of civil work that has to follow specific codes and methods of analysis can probably be done by AI. But no one is building a bridge or building without a licensed human signing off on the paperwork, and I imagine the liability around signing off on AI work will raise salaries of licensed engineers, and human engineers involved in the process.

But one senior engineer can feed AI some equations and boundary conditions and the AI can script tools for running solutions and cases very fast. One engineer can solve more and move faster than before. The engineer becomes more of a lead, writing instruction to AI on approach and technical requirements, and AI will write code to solve it. The engineer needs enough experience to debug the code and validate the outputs with test cases. But now the engineer is enabled to own a lot more of the design.

From my own experience and that of coworkers, AI still struggles mightily with basic undergrad engineering questions. You can't just give it a simple question and receive reliable and correct answers. You have to be extremely explicit to the point, AI isn't really doing the thinking work, it's just translating your instructions into code syntax and functions.

---

As maverick from top gun said- don't think just do. If your job can be replaced by AI you already are not providing much value

bmks270 said:

---

hph6203 said:

---

Would be stunning if there were enough inefficiency on Polymarket to average ~35% returns per trade over the course of 200 trades. If that were possible it would be wiped by people doing the same thing.

---

No, betting sites just ban you if your win rate is too high.

If AI is so good, then AI models will be used to set the odds.

---

Polymarket is not a bookmaker, they don't operate as the house, they are a market administrator. A market is created with a seed (normally not from Polymarket), the seed creates tokens, the tokens are purchased by traders, the token prices shift based upon demand for each token in the pool. Mismatches in demand for tokens are settled by arbitragers who either create additional tokens, and sell them, or buy both sides of the market and destroy them. They can make low single digit/sub 1% profits performing that task. It operates closer to an options market than Vegas.

Polymarket does not care about the success/profitability of users on their platform. Not only is it not detrimental to them, but their business depends on high volumes of profitable trades to make money. Their fees are accumulated by closed positions on profitable trades. They take 2% of profitable trades. They look at a graph like that and think "Awesome! Lots of trades, lots of profit, lots of money for us."

What that guy is claiming to do is scrape Polymarket posted markets and then scraping the internet for relevant information to the market, and buying a position prior to that novel information being materialized into the pricing split in the market. He's essentially claiming he's faster than the rest of the market to the tune of a profit margin of 36% per trade on the second day, when his profit per trade was only ~20.3% on the first day, using an agent he created in under a week where very intelligent people are already targeting inefficiency opportunities.

Return in 3 years and see how you feel about this statement.

I think what he does not get is that year by year, more and more jobs are not going to be providing any value!


The office lady taking shorthand was an invaluable employee. Until she wasn't.


All these people thinking that AI is not going to affect their job market have their head in the sand

hph6203 said:

---

bmks270 said:

---

hph6203 said:

---

Would be stunning if there were enough inefficiency on Polymarket to average ~35% returns per trade over the course of 200 trades. If that were possible it would be wiped by people doing the same thing.

---

No, betting sites just ban you if your win rate is too high.

If AI is so good, then AI models will be used to set the odds.

---

Polymarket is not a bookmaker, they don't operate as the house, they are a market administrator. A market is created with a seed (normally not from Polymarket), the seed creates tokens, the tokens are purchased by traders, the token prices shift based upon demand for each token in the pool. Mismatches in demand for tokens are settled by arbitragers who either create additional tokens, and sell them, or buy both sides of the market and destroy them. They can make low single digit/sub 1% profits performing that task. It operates closer to an options market than Vegas.

Polymarket does not care about the success/profitability of users on their platform. Not only is it not detrimental to them, but their business depends on high volumes of profitable trades to make money. Their fees are accumulated by closed positions on profitable trades. They take 2% of profitable trades. They look at a graph like that and think "Awesome! Lots of trades, lots of profit, lots of money for us."

What that guy is claiming to do is scrape Polymarket posted markets and then scraping the internet for relevant information to the market, and buying a position prior to that novel information being materialized into the pricing split in the market. He's essentially claiming he's faster than the rest of the market to the tune of a profit margin of 36% per trade on the second day, when his profit per trade was only ~20.3% on the first day, using an agent he created in under a week where very intelligent people are already targeting inefficiency opportunities.

---

So, what you're saying is, it's gambling.

Can AI replace realtors? I swear everytime I've used one, I've done more work and negotiated a better deal than they have.

The thing with this is that based upon CURRENT technology, I can extrapolate and see how it could replace about 95% of white collar jobs. That alone is cause for concern.

That will lead to hundreds of thousands or more of displaced jobs. May not happen but it is a very real possibility.

And that possibility grows each day. That's the scary part.

For the average person in the market. The collective knowledge of the market is largely accurately pricing probabilities. The edge occurs in the highly knowledgeable (or insider trader) and the knowledge of the market, which isn't a 35% margin per trade. If he had posted a $10,000 account with 1000 trades over a week with an average trade size of $50 and a 3-5% profit per trade I'd be way more inclined to believe it, but still skeptical. It just wouldn't be so quickly dismissed as fake.

I think American is attempting to use AI at DFW.

Retired civil engineer here. Career in land development, commercial and residential projects. 25 years ago the hot 'cost savings' pitch was outsourcing computer drafting and 'simple' design work to India for overnight turn around. Major failure at the firm I was in.

Now some folks who have never done it tell me that AI can 'do' all or even some of the human tasks civil engineers do to take a client's concept from start to finish. Snake Oil sales guys are with us always. You think a little parking lot and commercial site can be 'done' by AI? Tell me what 'done' is, first?

It ain't easy when it ain't.

reineraggie09 said:

---

Can AI replace realtors? I swear everytime I've used one, I've done more work and negotiated a better deal than they have.

---

That's because the field attracts people that want to turn a fast and easy buck. They aren't the hardest workers and brightest bulbs in the workforce.