AI agents wrong ~70% of time: Carnegie Mellon study
Technology
269
Beiträge
106
Kommentatoren
64
Aufrufe
-
and doesn't need to be exactly right
What kind of tasks do you consider that don't need to be exactly right?
Make a basic HTML template. I'll be changing it up anyway.
-
I'd just like to point out that, from the perspective of somebody watching AI develop for the past 10 years, completing 30% of automated tasks successfully is pretty good! Ten years ago they could not do this at all. Overlooking all the other issues with AI, I think we are all irritated with the AI hype people for saying things like they can be right 100% of the time -- Amazon's new CEO actually said they would be able to achieve 100% accuracy this year, lmao. But being able to do 30% of tasks successfully is already useful.
Please stop.
-
This post did not contain any content.
Color me surprised
-
and doesn't need to be exactly right
What kind of tasks do you consider that don't need to be exactly right?
Things that are inspiration or for approximations. Layout examples, possible correlations between data sets that need coincidence to be filtered out, estimating time lines, and basically anything that is close enough for a human to take the output and then do something with it.
For example, if you put in a list of ingredients it can spit out recipes that may or may not be what you want, but it can be an inspiration. Taking the output and cooking without any review and consideration would be risky.
-
Ok what about tech journalists who produced articles with those misunderstandings. Surely they know better yet still produce articles like this. But also people who care enough about this topic to post these articles usually I assume know better yet still spread this crap
Check out Ed Zitron's angry reporting on Tech journalists fawning over this garbage and reporting on it uncritically. He has a newsletter and a podcast.
-
At least AI won't fire you.
It kinda does when you ask it something it doesn't like.
-
and doesn't need to be exactly right
What kind of tasks do you consider that don't need to be exactly right?
Most. I've used ChatGPT to sketch an outline of a document, reformulate accomplishments into review bullets, rephrase a task I didnt understand, and similar stuff. None of it needed to be anywhere near perfect or complete.
Edit: and my favorite, "what's the word for..."
-
Please stop.
I'm not claiming that the use of AI is ethical. If you want to fight back you have to take it seriously though.
-
I'm not claiming that the use of AI is ethical. If you want to fight back you have to take it seriously though.
It cant do 30% of tasks vorrectly. It can do tasks correctly as much as 30% of the time, and since it's llm shit you know those numbers have been more massaged than any human in history has ever been.
-
It cant do 30% of tasks vorrectly. It can do tasks correctly as much as 30% of the time, and since it's llm shit you know those numbers have been more massaged than any human in history has ever been.
I meant the latter, not "it can do 30% of tasks correctly 100% of the time."
-
I meant the latter, not "it can do 30% of tasks correctly 100% of the time."
You get how that's fucking useless, generally?
-
... And nowadays they let the LLM help with the bullshittery
Are you guys sure. The media seems to be where a lot of LLM hate originates.
-
I'd just like to point out that, from the perspective of somebody watching AI develop for the past 10 years, completing 30% of automated tasks successfully is pretty good! Ten years ago they could not do this at all. Overlooking all the other issues with AI, I think we are all irritated with the AI hype people for saying things like they can be right 100% of the time -- Amazon's new CEO actually said they would be able to achieve 100% accuracy this year, lmao. But being able to do 30% of tasks successfully is already useful.
It doesn't matter if you need a human to review. AI has no way distinguishing between success and failure. Either way a human will have to review 100% of those tasks.
-
I’m sorry as an AI I cannot physically color you shocked. I can help you with AWS services and questions.
How do I set up event driven document ingestion from OneDrive located on an Azure tenant to Amazon DocumentDB? Ingestion must be near-realtime, durable, and have some form of DLQ.
-
Are you guys sure. The media seems to be where a lot of LLM hate originates.
Whatever gets ad views
-
You get how that's fucking useless, generally?
yes, that's generally useless. It should not be shoved down people's throats. 30% accuracy still has its uses, especially if the result can be programmatically verified.
-
It doesn't matter if you need a human to review. AI has no way distinguishing between success and failure. Either way a human will have to review 100% of those tasks.
Right, so this is really only useful in cases where either it's vastly easier to verify an answer than posit one, or if a conventional program can verify the result of the AI's output.
-
How do I set up event driven document ingestion from OneDrive located on an Azure tenant to Amazon DocumentDB? Ingestion must be near-realtime, durable, and have some form of DLQ.
I see you mention Azure and will assume you’re doing a one time migration.
Start by moving everything from OneDrive to S3. As an AI I’m told that bitches love S3. From there you can subscribe to create events on buckets and add events to an SQS queue. Here you can enable a DLQ for failed events.
From there add a Lambda to listen for SQS events. You should enable provisioned concurrency for speed, the ability for AWS to bill you more, and so that you can have a dandy of a time figuring out why an old version of your lambda is still running even though you deployed the latest version and everything telling you that creating a new ID for the lambda each time to fix it fucking lies.
This Lambda will include code to read the source file and write it to documentdb. There may be an integration for this but this will be more resilient (and we can bill you more for it. )
Would you like to see sample CDK code? Tough shit because all I can do is assist with questions on AWS services.
-
yes, that's generally useless. It should not be shoved down people's throats. 30% accuracy still has its uses, especially if the result can be programmatically verified.
Less broadly useful than 20 tons of mixed texture human shit, and more ecologically devastatimg.
-
Less broadly useful than 20 tons of mixed texture human shit, and more ecologically devastatimg.
Are you just trolling or do you seriously not understand how something which can do a task correctly with 30% reliability can be made useful if the result can be automatically verified.
