- Joined
- Nov 21, 2023
- Messages
- 1,053
- Reaction score
- 1,647
- Points
- 113
I'm going to step out on a limb and say you probably have a horseless carriage too ;-)This thread had me intrigued, so I thought I would check it out for myself. My next beer is a Saison that I make every summer, so I decided to compare my beer with an AI version. I think the answer is solid enough and like any recipe I see I can nit-pick a bit (let's face it, Flaked Oats will not improve head retention). The main sources were from Craft Beer and Brewing, The Mad Fermentationist, and Brew Your Own.
"I need a recipe for a Saison with Spelt malt"
AI Overview
This classic spelt saison recipe utilizes spelt malt to yield a dry, effervescent, farmhouse-style ale with a rustic, slightly nutty character. [1, 2]
Batch Details
- Batch Size: 5 Gallons (19 L)
- Original Gravity (OG): 1.050 - 1.054
- Final Gravity (FG): 1.002 - 1.006
- ABV: ~6.0% - 6.5%
- IBUs: ~25 - 30 [1, 2]
Ingredients
Grains & Fermentables [1, 2]
Hops Schedule
- 6.5 lbs (2.95 kg) Belgian Pilsner Malt
- 3.0 lbs (1.35 kg) Spelt Malt
- 1.0 lb (454 g) Flaked Oats (for head retention and soft mouthfeel)
- 0.5 lbs (227 g) Vienna Malt (for a touch of color and complexity)
- 0.5 lbs (227 g) Table Sugar (added during the boil to boost attenuation) [1, 2, 3]
Yeast
- 1 oz (28 g) Styrian Goldings (60 min) -> ~19 IBUs
- 1 oz (28 g) Styrian Goldings or Saaz (15 min) -> ~7 IBUs
- 0.5 oz (14 g) Styrian Goldings or Saaz (Flameout/Whirlpool) -> ~2 IBUs [1, 2, 3]
- White Labs WLP565 (Belgian Saison), Wyeast 3724 (Farmhouse Ale), or Fermentis SafAle BE-134 [1]
Brewing Instructions
1. Mash
2. Boil
- Malt Handling: Spelt has a strong, sticky husk. Rice hulls (0.5 lbs) are highly recommended to prevent a stuck mash.
- Step Mash: Mash in at 122°F (50°C) for a 10-minute protein rest.
- Conversion: Raise to 149°F (65°C) and rest for 60 minutes.
- Mash Out: Raise to 170°F (77°C), hold for 10 minutes, and vorlauf. [1, 2, 3, 4, 5]
3. Fermentation & Conditioning
- Boil Duration: 90 minutes.
- Sugar Addition: Add the table sugar with 15 minutes left in the boil, ensuring it dissolves completely. [1]
- Pitching: Pitch yeast at 68°F (20°C) and aerate well.
- Temperature Control: Start fermentation around 68°F, and allow it to free-rise to 75°F - 80°F (24°C - 27°C). Saison yeasts like warm temperatures to finish completely dry.
- Aging: Bottle or keg to a high carbonation level (~2.5 - 3.0 volumes) to enhance the spritzy, dry profile. [1, 2, 3, 4, 5]
<kidding guys !>
From what I've seen from the consumer / non-corporate agents, and particularly the later models, this was a really basic conversation. Try feeding in your equipment profile, flavor profile you're looking for or something similar and let it work out what you want to get to via q&a.
I haven't specifically tried hop combinations or hop substitution q&a yet but I suspect it would work reasonably well at that.
All of this said, I kind of liken AI agents to graphical user interfaces; GUI's I saw an implementation that intended to "graphicalize" mainframe computer functions. My response was "Great, that's all we need is to dumb down access to critical functions so any [computer] operator can IPL from the wrong volume". The one being demo'd didn't last too long, but other, similar efforts to 'front' complex mainframe procedures did keep advancing, much like the server farm platforms respective GUIs to front command line functionality.
The net result: Companies began hiring less skilled Sysadmins and Systems Programmers who could rely on the GUI and use tech support to bail them out. Big companies. With lead system programmers that if you were being honest, you wouldn't let wind your watch, much less be in charge of a multi-billion dollar enterprise. I'll paraphrase one of my favorite examples from about 3 years ago:
"I only know 2 things about your product, reboot and reload the config. Tell me what I'm gonna do to fix this problem".( said with a THICK New York accent) That is a direct quote, absent the name of the product. This is the quality of staff leading some of the computer rooms responsible for safeguarding _OUR_ customer data.
Just like the GUI before it, AI will dumb down more complex programming, repetitive tasks to a very low level. It will bring a lot of tasks to the masses who really should not have it because once something breaks, they won't be able to practically fix it.
So, AI can be a cool tool to use, I trust hands on experience more.


