4 ms·

Yikes! I can only imagine an LLM hallucinating my portfolio to zero.

Incipient 1y ago
It may not hallucinate yours to zero, or mine, or franks, or Mary's, but at some point it will do it to someone. That's the issue I have with these approaches *at scale*. I'm sure one day we'll get 100.00% reliable outputs such that autonomous agents can do this, however that's not today.
- malfist 1y ago
  We have those today, it's just not called AI, it's called things like idempotent APIs
- sgt101 1y ago
  Yeah - feels like lunacy to me. I love the way that they get to the end of the article and then write that it doesn't really work!
  - sgt101 1y ago
    I remember that John McCarthy used to condemn AI research as being a series of "look mom no hands" demos, but now we've got to "look mom I face planted". This work would be valid if it did things like : - Describe an intelligent effect rather than "invoked an API" - so something like the system evaluating the portfolio and the market and deciding what should be sold or offering buying opportunities based on some inference (this is not a great example for many reasons). - Measure and report the systems performance. The write-up says that the LLM fails, ok... how often? - Describe the failure cases, provide some theory as to why some things succeeded and others didn't. - Provide a way forward. What's next? Without doing these things this work isn't helpful and is part of the AI/Agent/MCP hype. Basically it's Bored Apes for AI.
- bboygravity 1y ago
  The way this is done at scale is legalized front-running. You buy order flow from "free" retail brokers like Robin Hood (as a market maker) and then literally front run ALL trades. You then tell everybody you're giving them the best price (also the only price), lol. In other words: legalized crime.
  - diggan 1y ago
    > legalized crime Hmm? Either it's legal, and not a crime. Or it's illegal, and therefore a crime. What you mean is "legalized crime" exactly? Things considered legal but unethical for whatever reason?
  - drexlspivey 1y ago
    You need microsecond latency for HFT, certainly not enough time to consult an LLM
  - MOARDONGZPLZ 1y ago
    What does this mean? Sorry, not up on stock lingo. Is this effectively like having the results of a race a minute early or something?
    - rco8786 1y ago
      Yea that's a decent analogy. You see the incoming trades before they are actually executed, and you can react accordingly.
  - dist-epoch 1y ago
    If somebody wants to buy an orange for $2 and I sell them one for $1.98, that's "legalized crime"? I don't think you understand how this works. It's not about front-running them, it's about that they have no clue what they are doing, so it's extremely unlikely their trade will be profitable, thus you gain by being the other side. That $0.02 of price improvement is the price you pay for the privilege of giving a better price to clueless people.
    - 1y ago

It may not hallucinate yours to zero, or mine, or franks, or Mary's, but at some point it will do it to someone. That's the issue I have with these approaches *at scale*. I'm sure one day we'll get 100.00% reliable outputs such that autonomous agents can do this, however that's not today.

malfist 1y ago
We have those today, it's just not called AI, it's called things like idempotent APIs
sgt101 1y ago
Yeah - feels like lunacy to me. I love the way that they get to the end of the article and then write that it doesn't really work!
- sgt101 1y ago
  I remember that John McCarthy used to condemn AI research as being a series of "look mom no hands" demos, but now we've got to "look mom I face planted". This work would be valid if it did things like : - Describe an intelligent effect rather than "invoked an API" - so something like the system evaluating the portfolio and the market and deciding what should be sold or offering buying opportunities based on some inference (this is not a great example for many reasons). - Measure and report the systems performance. The write-up says that the LLM fails, ok... how often? - Describe the failure cases, provide some theory as to why some things succeeded and others didn't. - Provide a way forward. What's next? Without doing these things this work isn't helpful and is part of the AI/Agent/MCP hype. Basically it's Bored Apes for AI.
bboygravity 1y ago
The way this is done at scale is legalized front-running. You buy order flow from "free" retail brokers like Robin Hood (as a market maker) and then literally front run ALL trades. You then tell everybody you're giving them the best price (also the only price), lol. In other words: legalized crime.
- diggan 1y ago
  > legalized crime Hmm? Either it's legal, and not a crime. Or it's illegal, and therefore a crime. What you mean is "legalized crime" exactly? Things considered legal but unethical for whatever reason?
- drexlspivey 1y ago
  You need microsecond latency for HFT, certainly not enough time to consult an LLM
- MOARDONGZPLZ 1y ago
  What does this mean? Sorry, not up on stock lingo. Is this effectively like having the results of a race a minute early or something?
  - rco8786 1y ago
    Yea that's a decent analogy. You see the incoming trades before they are actually executed, and you can react accordingly.
- dist-epoch 1y ago
  If somebody wants to buy an orange for $2 and I sell them one for $1.98, that's "legalized crime"? I don't think you understand how this works. It's not about front-running them, it's about that they have no clue what they are doing, so it's extremely unlikely their trade will be profitable, thus you gain by being the other side. That $0.02 of price improvement is the price you pay for the privilege of giving a better price to clueless people.
  - 1y ago