ive been using AI a lot lately, realized just now that it can dig a lot deeper than a standard google search to find information on commercial beers.
Just got:
Beer Variant
ABV
Estimated Original Gravity (OG)
Estimated Final Gravity (FG)
Chimay Blue
9.0%
1.080 – 1.090
1.010 – 1.012
Chimay White
8.0%
1.070 – 1.075
1.008 – 1.012
Chimay Red
7.0%
1.061 – 1.064
1.008 – 1.010
Chimay Gold
4.8%
1.044 – 1.048
1.006 – 1.008
and just now as i post this i searched and got:
Beer Variant
ABV
Estimated Original Gravity (OG)
Estimated Final Gravity (FG)
Westmalle Tripel
9.5%
1.080 – 1.084
1.008 – 1.012
Westmalle Dubbel
7.0%
1.063 – 1.065
1.008 – 1.012
Westmalle Extra
4.8%
1.044 – 1.048
1.008 – 1.011
of course you have learned that not everything you read online is true - but you can ask “how did you get this info? / show me the primary documents you used.” and dig deeper.
i will probably be playing around with this for now, but i am making this thread as honestly - LLM info gathering is going to be a great update to guessed at info from the internet of ~2005 to 2025
i havent read BLAM in a long while. as well, looking at these numbers they are what one would expect, no surprises.
I just wanted to focus on using AI in general to essentially dig deeper into documentation out there that otherwise might be undiscovered by forum users.
i dont want to just fill this thread with copied and pasted stuff but another example:
"Core Technical Specifications
Based on the BYO publication and Kimmich’s shared data, the technical vitals for a standard 5.5-gallon batch are as follows:
Original Gravity (OG): 1.076.
Final Gravity (FG): 1.014.
Alcohol by Volume (ABV): 8% (Standard production target).
International Bitterness Units (IBU): 100+ (Estimated between 100 and 120+).
Kimmich has emphasized that technique and water chemistry are more vital than the specific hop percentages. [1, 2]
Grain Bill: Primarily British 2-row (often Pearl or Maris Otter) and a small amount of Caravienne malt. A significant portion of fermentables (approx. 1 lb) comes from turbinado sugar to ensure a dry finish despite the high OG.
Hop Profile: Uses a complex blend of roughly six varieties, including Simcoe, Centennial, Cascade, Apollo, Chinook, and Columbus.
Yeast: The proprietary “Conan” strain, now commercially available as “Vermont Ale” or “Burlington Ale” yeast.
Water Profile: Known for a very high sulfate-to-chloride ratio (upwards of 237–760 ppm sulfate) to accentuate the sharp hop bitterness"
again - im a big believer in following statements back to their source - so again you simply ask “show me the primary documents for this info - MLA format” or whatever you want.
i just think its neat!
edit: and further i did just ask it to compare that to blam - looks like the lower numbers in the ranges are the BLAM ones.
it includes BLAM basically as a source
edit 2: perhaps we could even make a live spreadsheet a la DMtaylor’s yeast one ?? but for commercial beers
I’m using AI a lot for my salt additions. It’s a lot easier than the calculators out there and so far, even though it’s significantly different from the calculators, I’ve found it spot on for the few recipes I’ve brewed using it so far.
As a point of clarification, what do you mean by AI being significantly different from the calculators? As in, it is giving you very different mineral addition amounts?
Yes, the amounts, and in two recipes it suggested adding a very small amount of epsom, which I don’t normally do. I use BF for my recipes and find myself obsessively trying to tweak the salt amounts to change the delta numbers from red to green, or at least get close. I know part of it is my compulsive behavior. With AI, I tell it all aspects of the recipe and it spits out the exact amounts I need to add.
FWIW, I use epsom salt pretty regularly when adjusting our tap water, but I also keep a close eye on magnesium levels. The local water supply has a ton of calcium, so gypsum can tip things too far!
But I agree that epsom salt is not my default when building from RO!
i was having off tang issues with batch after batch of beer. i did a strip down of my at-that-time standard basic practices which often included adding USP epsom salts even if in small amounts. i removed a bunch of features i was doing such as using distilled (not RO - i incorrectly assumed it was the same) water and it went away. havent tried to adjust my mg levels since and do not have that problem anymore.
at a minimum it wasnt required for my beers.
re: AI suggestions - yes there are going to be different ways of using it - ie. asking it to brew a beer based on subjective data vs. requesting it to gather primary documents for you and summarizing that data.
THOUGH in truth i just used AI to formulate a hop schedule for a bitter:
30 IBU magnum
30g EKG at 20min
14g el dorado at 10
14 g mosaic, 14g simcoe at 77C whirlpool
dryhop with 1/2oz each mosaic el dorado and simcoe for a few days cold
I agree that AI tools are usually good for research and aggregating data. And I also would only rely on it if it will cite its sources so that I can verify the information. But if it saves me the time I’d spend searching, I’m good with that. (Although, I’ve discovered some interesting things in the past while searching, so I may miss that.)
To Denny’s point, I don’t think that this removes the creative part, unless you choose that route. I usually do research before creating a recipe, and using AI as my search engine just makes that faster so that I get to the creative part sooner. I don’t have enough experience / knowledge to skip the research step.
Using brewing software is basically cheating, it takes the art and science out of sitting down and understanding the ppg per lb of grain. Personally i prefer to keep my brewnotes on pen and paper, doing the equations each time to calculate expected OG and IBU.
@Skeeter86
the term itself AI is the misleading part. I see you do understand it, but more for anyone reading this - LLM is a better term. I think of the relationship as like a search engine vs. using the dewey decimal system at a physical library.
you can always ask it to cite its sources and you are free to check out each one, criticize its results and tell it to reformulate, explore something new.
Of course I didn’t just say “make a recipe for me that tastes like X”, it took several generations of discussion to come to it.
I use AI now at work to simply remove monotonous excel/copy and pasting tasks asking it to analyze some very raw text and extract and organize what i need.
i guess you could say aids in achieving “simple homebrewing”
I occasionally use ChatGPT to “design” recipes for styles that I have less familiarity with, ingredients-wise. I then filter the results through the 30 years’ experience with ingredients that’s in my brain. The recipes suggested by ChatGPT are just baselines, but they can be useful starting points.
LLMs (and I agree they should be called this vs. AI) can be useful tools provided one has enough capacity for critical thinking to view their results through a skeptical lens. But critical thinking and skepticism are in short supply these days…sigh.
you may think im AI mental here, but i see this as a really powerful tool that 99% of people do NOT see the application of yet.
ie. Homebrewing and beer drinking in general declining?
Let’s maximize the quality of our product by crowdsourcing (essentially) data to ensure most people are making the BEST homebrew and craft beer they can.
targeting a certain demographic? find what tastes they have in general and target that
Seems to me the overall beer market is cyclical over the years with various “fad” fermented products coming and going such as wine coolers ciders and now seltzers
Macro light lagers always sell well even though they have ups and downs too. My son who is now of age enjoys macros as well as IPA’s depending on the setting
I buy beer now and then but I am getting burnt out on “juicy” this and “milkshake” that. Virtually all are about 6 to 7% and while brewed by different breweries they are virtually the same beer.
Hence my propensity to buy well made lagers and brew as well as buy beer 4% and under. But that is just me
“quality” is a relative term, im saying the profiles that people actually want + the highest quality expressions of existing beers and future ones
@HEUBrewer yes, i know its said alot but i do feel it may be regional as to which areas are worse hit by the flood of low quality hazy/juicy everything. most places that have beer they will have all shelves filled with macro-garbage NAILs/“Try Local!” generally poor quality boring cream ale/“CRAAFFTT” lager/session IPA/Hazy/Juicy IPA and then some wider distributed quality craft brewed hazy/juicy/west coast IPAs + a few random sours that are not great.
None of this scratches my itch except occasionally getting a few select IPAs by the better craft breweries in my province. but imho hazy and juicy IPAs are one of the least sessionable beers i can imagine.
I typically only buy craft lagers that are brewed by lager breweries. We have a fabulous one here in the Chicago area, Goldfinger, incredible beer especially on draught. There are exceptions such as Bearded Iris in Nashville but I have found ale breweries can’t do European lagers very well.
Also even if there are exceptional breweries you have to be vigilant on bottling/canning dates. Even our most popular liquor (beer) store have old cans which impacts flavor. I have left with nothing many times because of old beers on the shelves. If it’s more than 2 months from canning/bottling I am not buying it
