AI analysis and tier list of peptide vendors?
-
@stevepep I see what people euphemistically refer to as hallucinations all the time. In the peptide space, it’s awful because there is a ton of bad information that gets spread widely so the model assumes frequency = truth when it is just many sources repeating one piece of bad information.
I work in financial derivatives and as you say, it can be really good for coding, but if you are trying to get information in a field without a lot of accurate information, it fills in the blanks.
One of my funniest AI stories was a real estate search where Copilot kept returning a listing for an empty lot while insisting that there was a 4 bedroom house on the property. And every time i tried to correct it, it gave me the old, “You’re right to call me out on that, here is a better list…” and then gave me the exact same listing for an empty lot.
-
@stevepep I see what people euphemistically refer to as hallucinations all the time. In the peptide space, it’s awful because there is a ton of bad information that gets spread widely so the model assumes frequency = truth when it is just many sources repeating one piece of bad information.
I work in financial derivatives and as you say, it can be really good for coding, but if you are trying to get information in a field without a lot of accurate information, it fills in the blanks.
One of my funniest AI stories was a real estate search where Copilot kept returning a listing for an empty lot while insisting that there was a 4 bedroom house on the property. And every time i tried to correct it, it gave me the old, “You’re right to call me out on that, here is a better list…” and then gave me the exact same listing for an empty lot.
@ResearchCat my favorite hallucination is when people use AI callings and the person asks can you give me my tracking number when it can’t find it it just makes it up.
Completely useless in this context there is no reliable data for it to rely on.
-
Speaking of peptide vendor analysis, has anyone heard of https://titratelab.com? I saw this posted in another forum I belong to.
-
I've had a couple of unknown to me vendors recommended lately, was wondering if anyone here has heard of them (didn't get any hits in the forum search engine);
-
Precision Research Labs (aka PRL), I was told that test every vial for endotoxin, microbial, heavy metal, and fentanyl contamination
-
RCHQ, close friend has been reporting good things about them
-
Kimera Chems, several acquaintances at local gym have been recommending them
Here's the ChatGPT first breakdown on the three:
Updated after providing screenshot of PRL's new COA claims:
I would appreciate any insights if folks here have had any good or bad experiences to report regarding these 3 vendors.
-
-
Boo! Please don’t make me read Ai slop.
-
Boo! Please don’t make me read Ai slop.
@ResearchCat I get why people do it but in these cases you are just getting what’s available on the internet and other unrestricted areas so it’s mostly slop like you said. Crap in, Crap out. In peptide stuff I just send it to an area like the fda website to search for certain things it’s useful in that capacity.
-
I don’t want to rant, but the issue is the same as at work. People send me very long AI-generated emails that have a low information ratio and take a lot of time to read to determine no useful information is there. I encourage people to use AI selectively and with discretion with respect to the results, but - no offense to the poster - asking me to read the results of whatever prompt you entered is wasting limited time and not getting you the answer you’re looking for.
It also adds to the global self-reinforcing database of garbage AI said is true which AI will now have available to use as the basis for telling you again.
“Have any of you heard of these guys and are they any good?” Is a short and effective human-targeted prompt that will result in better results faster.
And without the aggro. -
I don’t want to rant, but the issue is the same as at work. People send me very long AI-generated emails that have a low information ratio and take a lot of time to read to determine no useful information is there. I encourage people to use AI selectively and with discretion with respect to the results, but - no offense to the poster - asking me to read the results of whatever prompt you entered is wasting limited time and not getting you the answer you’re looking for.
It also adds to the global self-reinforcing database of garbage AI said is true which AI will now have available to use as the basis for telling you again.
“Have any of you heard of these guys and are they any good?” Is a short and effective human-targeted prompt that will result in better results faster.
And without the aggro.@ResearchCat
I am getting an influx of people in my line of work of people submitting AI generated stuff. 99% of the time it is all wrong, and like you said low amount of info with a high amount of fluff...
I actually use AI even though I am a Techno Peasant, but I really only use it as a code generator. It is better than the code generators we had in the past, but comes with its own set of new problems that we didn't have with the old code generators, like guessing... -
One of the funny things about using AI to generate code is that in the past, you needed a BA who understood what the requirements were and a developer who understands what needs to be built and how to do it, and most of the time it doesn’t come out right. Now, you still need someone who understands what the requirements are, but fewer and fewer people can tell you if it is actually doing what it needs to do.
-
I don’t want to rant, but the issue is the same as at work. People send me very long AI-generated emails that have a low information ratio and take a lot of time to read to determine no useful information is there. I encourage people to use AI selectively and with discretion with respect to the results, but - no offense to the poster - asking me to read the results of whatever prompt you entered is wasting limited time and not getting you the answer you’re looking for.
It also adds to the global self-reinforcing database of garbage AI said is true which AI will now have available to use as the basis for telling you again.
“Have any of you heard of these guys and are they any good?” Is a short and effective human-targeted prompt that will result in better results faster.
And without the aggro.@ResearchCat I could rant about employees spitting out AI generated garbage. I had to put an end to it you get pages of nonsense. It won’t say it doesn’t know so it spits out crap half the time. I came close to letting someone go last week over a plan I asked for that was clearly AI and a waste of my time. It’s not ready for prime time but people rely on it like it is.
-
AI slop. No thank you. It really gets fun when you have other AI models checking the work.
-
One of the funny things about using AI to generate code is that in the past, you needed a BA who understood what the requirements were and a developer who understands what needs to be built and how to do it, and most of the time it doesn’t come out right. Now, you still need someone who understands what the requirements are, but fewer and fewer people can tell you if it is actually doing what it needs to do.
@ResearchCat
What I see with AI code generator, it does not use best practices. By example, it will put the same data In multiple places and tables, and does not always use unique identifiers. That is rule number one in object oriented programming. Most people will be able it to build apps and stuff that does not really matter and isn't that big. I built my peptracker app with it.
Now my case management / email client / accounting / field specific software that I am currently building requires a knowledge and background in computer programming and network security. I am 3 months into it and have another 4 months to go...
AI code generators have made more mistakes in building the code for that than I can count. It has made me realize how truly limited AI currently is. Is is a far way off... and that is why i don't trust any report it gives. -
I don’t want to rant, but the issue is the same as at work. People send me very long AI-generated emails that have a low information ratio and take a lot of time to read to determine no useful information is there. I encourage people to use AI selectively and with discretion with respect to the results, but - no offense to the poster - asking me to read the results of whatever prompt you entered is wasting limited time and not getting you the answer you’re looking for.
It also adds to the global self-reinforcing database of garbage AI said is true which AI will now have available to use as the basis for telling you again.
“Have any of you heard of these guys and are they any good?” Is a short and effective human-targeted prompt that will result in better results faster.
And without the aggro. -
Boo! Please don’t make me read Ai slop.
-
@ResearchCat
Not a huge fan of AI myself but after failing to turn anything up on those three companies with the usual research methods... seemed like it was worth a shot.@MK2a1 perfectly valid. If neither google nor AI can give you anything useful, note there are plenty of well vetted and trusted vendors on here that many of us have been using for years.
Hello! It looks like you're interested in this conversation, but you don't have an account yet.
Getting fed up of having to scroll through the same posts each visit? When you register for an account, you'll always come back to exactly where you were before, and choose to be notified of new replies (either via email, or push notification). You'll also be able to save bookmarks and upvote posts to show your appreciation to other community members.
With your input, this post could be even better 💗
Register Login