These are things I've learned from experience that others might find helpful. Some of them are easy to miss for a while. (Also an exercise in "reality contains a surprising amount of detail"; I could probably have kept going for a while but needed to call it at some point.)
Baking
Oven thermostats are often miscalibrated enough to matter. If you're following existing recipes but find things often coming out overdone or underdone, you might consider buying an oven thermometer to check how miscalibrated your oven thermostat is. Unfortunately, oven thermometers are also often miscalibrated. Fortunately, they're not that expensive[1]. A friend of mine bought three from three different brands to check for inter-rater agreement. Note that ovens can end up at different temperatures in different locations within the oven[2], so ideally you want to place all three thermometers relatively closely together (but not touching) roughly around where you typically put the thing you're baking. (Also note that other factors can affect baking times, like altitude.)
You need to use mass measurements rather than volumetric measurements.For everything macro-scale, anyways - if a recipe asks for a teaspoon of vanilla extract, nobody will tell you how many grams that was supposed to be and there aren't that many available sources of variance. Much less the case for e.g. "cups of flour"! Flour in particular is highly compressible[3] and many recipes use highly unrealistic estimates of how many grams there are in a cup of flour[4], when telling you how many cups to use. Fortunately, the aforementioned friend also ran a Flour Measuring Science Party and walked away with a spreadsheet. And it turns out that you can hit 120 grams per cup if you carefully scoop the flour in with a fork, but if you just use the cup measure itself as the scoop you're more likely to end up at 140 grams, and deliberately packing it down can get you to 180. Which is to say: always use mass measurements when available. If a recipe website doesn't either default to mass measurements, or provide a toggle, that's a deeply negative sign about its quality. Relatedly...
Own a kitchen scale. You want one that lets you switch between different units and zero out the current weight. This one is pretty good; the ones that cost $12-15 are probably also pretty good.
Disposable shower caps are a huge improvement over saran wrap, when it comes to operations like "cover the bowl containing the bread dough while proofing". 95% reduction in effort. Something like these[5].
You can "prep" good bread dough in less than ten minutes. I recommend Zvi's transcription of the core recipe from The New Artisan Bread in Five Minutes a Day. I've never tried it with all-purpose flour; I recommend just purchasing bread flour. You may notice that the linked recipe uses volumetric measurements. Having made that recipe probably 20+ times now, and having a sense for how minor variations in flour/water ratios affect the dough, I can now provide you with reliable mass measurements instead, as well as other improvements and notes:
960g bread flour (I use King Arthur; the numbers might be different if you use a different bread flour with different a protein ratio)
670g water
1.5 tablespoons instant yeast (you can buy a pound for $5-10 and it keeps for many months in the fridge)
13.5 grams salt (1.5 tablespoons kosher salt / 0.75 tablespoons table salt)
Follow steps 1 - 3 in Zvi's post. You can speed up step 4 by heating your oven up to ~110F (or using a "warming mat") and tossing the covered dough in for ~75 minutes. You'll likely need to dial this process in yourself, so give it an extra 15 minutes at room temp the first couple times you do it to check how much more it rises - if it's a noticeable amount it probably wasn't quite there.
Zvi's step 5 says "Put it in the refrigerator and use as needed. It should be good for at least two weeks." I would go further and say that there's a noticeable improvement in bread quality after the first 12-24 hours of refrigeration, compared to making it immediately after the first rise. Fermentation will continue (slowly) over the coming days, which many but not all people regard as a positive.
Zvi then goes on to the actual baking instructions. Here are some additional notes of mine:
The recipe above makes basically exactly enough dough for two loaves if using typical 8x4/9x5 loaf pans.
Use a non-stick olive oil spray. Zvi suggests either greasing the pan with flour and butter, or using high-quality wax paper. I think I tried wax paper once and wasn't happy with the result. Butter and flour are annoying. Use a spray - hit all four sides and the bottom, rub a paper towel over the sides and bottom to ensure smooth distribution after spraying, and then dab away any excess oil that pools in a corner if you tilt the pan for a few seconds. This is much faster, you're much less likely to miss spots, it keeps the recipe vegan, and it serves the non-stick purpose better. There is of course a slight difference in taste/texture but it's basically a wash.
The second rise post-refrigeration also seems to be important for the bread quality. I've generally been much happier with a full second rise after taking the dough out of the fridge, than with no rise or a substantially shorter rise. So after greasing up the pan, putting in half the dough, and covering it up with a shower cap, do whatever you did for the first rise.
Zvi says of step 7: "Have a pan on the rack below the bread, and dump a cup of warm water onto that pan to generate steam. Again, slightly useful, not actually necessary. We mostly skip it." I've found it to help with the texture of the crust and it's a trivial amount of effort (you can use cold tap water, it's fine, just make sure your oven actually hits 450 after that). You probably only need like a third of a cup, not a full cup.
Let the bread rest for 15-30 minutes before cutting into it.The bread ends up gummy if you cut into it while it's still hot. This is bad. Some of my housemates disagree with this trade-off being worth it (they like it hot). But alas, I am the one making the bread.
Salted butter contains meaningful amounts of salt, but this is usually fine if you don't have unsalted butter. Table salt is about 40% sodium, so take the sodium quantity in the salted butter (don't forget to multiply the "per serving" by the number of "servings" in the quantity of butter you'll end up using) and multiply it by 2.5 to see how much you should reduce the quantity of "salt" (by mass) you add. A teaspoon of table salt is roughly 6 grams, so if a recipe calls for a stick of unsalted butter and a teaspoon of salt, and you only have a stick of salted[7] butter, just reduce it to two-thirds of a teaspoon[8]. Relatedly...
Pay attention to whether the recipe is asking for kosher salt or table/fine salt. Salt is one of those cursed ingredients that even the good recipe websites will generally only provide a volumetric measurement for, generally in teaspoons. The usual conversion ratio is to cut the volume of salt in half when going from kosher to table salt - that is, you should use half a teaspoon of table salt to substitute for a full teaspoon of kosher salt[9]. But, also, many recipes are not that sensitive to the exact amount of salt and if you overdo it by 20-30% you probably won't notice. (You might if you underdo it.)
Cakes are often a bad trade-off in terms of effort vs. reward. They often take dramatically more time than other things that people tend to like about as much, so if you're making a cake you should probably be trying to make something that doesn't have a relatively close substitute in the rest of dessert-space. If you just want "chocolate dessert" make these muffins. Other people might suggest brownies, but, ugh. There are some exceptions: with a bit of iteration and experimentation, Claude and I came up with a surprisingly good (and vegan!) chocolate cake recipe that you can probably prep in less than 30 minutes:
Dark Wacky Cake
INGREDIENTS
205 grams all-purpose flour
250 grams granulated sugar
60 grams dutch cocoa powder
1.3 teaspoons baking soda
0.5 teaspoons fine salt
0.8 teaspoons espresso powder
315 grams cold water or cold coffee
88 grams olive oil
1.3 tablespoons apple cider or white vinegar
1.3 teaspoons vanilla extract
STEPS
1. Preheat oven: Preheat oven to 350°F (175°C). Lightly grease a 9x9 inch pan or line with parchment paper.
2. Combine dry ingredients: In a medium bowl, whisk together all of the flour, sugar, cocoa powder, baking soda, salt, and espresso powder. Sift if the cocoa is lumpy—Double Dark tends to clump. Make sure the baking soda is evenly distributed throughout.
3. Combine wet ingredients: In a separate bowl or large measuring cup, combine cold water or cold coffee, olive oil, vinegar, and vanilla extract. Stir briefly to combine.
4. Mix batter: Pour the wet ingredients into the dry ingredients. Stir until just combined—you'll see some fizzing as the vinegar reacts with the baking soda. Don't overmix; a few small lumps are fine. The batter will be thinner than a typical cake batter.
5. Bake immediately: Pour the batter into your prepared pan right away—the leavening reaction is happening now. Bake for 28-32 minutes, until a toothpick inserted in the center comes out with moist crumbs (not wet batter, not bone dry).
6. Cool: Let the cake cool in the pan for 10 minutes, then serve directly from the pan or turn out onto a rack. Top with powdered sugar, ganache, or eat plain.
Relatedly, modern frontier LLMs[10] are surprisingly useful cooking assistants:
They can provide reasonable recipes for basically everything that already has a name. You can sometimes get slightly improved results by asking them to fetch existing recipes for [baked good] from the recipe websites it knows to be high-quality, comparing them, and then giving you a synthesized recipe based on first-principles reasoning about why those recipes might have differed[11].
They also know what the "good" recipe websites are, at least for the domains that I've tried asking them for recommendations. Recipes from the good recipe websites will often be better than the generic good-enough recipe you can get from the LLM. I'm partial to Smitten Kitchen.
They seem quite good at suggesting safe/low-impact ingredient substitutions, though I've only tried this like five times.
Recipe websites are deeply unreliable for estimates of "prep time". Generally their estimates are dramatic underestimates for total time from "getting off the couch" to "thing goes in oven", even if you're an experienced home baker. Frontier LLMs also often make mistakes like this, at least with the kind of naive prompting that I've tried so far[12].
Thanks to Drake Thomas for getting me into baking, introducing me to Smitten Kitchen, buying three oven thermometers, and hosting a Flour Measuring Science Party.
To prep! Your fastest end-to-end time for actually having bread that you could even in theory eat is something like 2 hours, and that'd be cutting a lot of corners.
This is apparently only true for the Diamond Crystal brand of kosher salt; you multiply by 0.75 rather than by 0.5 if translating from Morton's kosher salt. But apparently most recipes assume Diamond Crystal, so "cut it in half" is usually correct. The additional facts in this footnote I learned from LLMs while writing this post, so consider taking them with a grain of salt.
I haven't tried anything more clever than "Give me recipes following [constraint x] that take less than 45 minutes", or "How much prep time will [recipe] take?"
These are things I've learned from experience that others might find helpful. Some of them are easy to miss for a while. (Also an exercise in "reality contains a surprising amount of detail"; I could probably have kept going for a while but needed to call it at some point.)
Baking
Oven thermostats are often miscalibrated enough to matter. If you're following existing recipes but find things often coming out overdone or underdone, you might consider buying an oven thermometer to check how miscalibrated your oven thermostat is. Unfortunately, oven thermometers are also often miscalibrated. Fortunately, they're not that expensive[1]. A friend of mine bought three from three different brands to check for inter-rater agreement. Note that ovens can end up at different temperatures in different locations within the oven[2], so ideally you want to place all three thermometers relatively closely together (but not touching) roughly around where you typically put the thing you're baking. (Also note that other factors can affect baking times, like altitude.)
You need to use mass measurements rather than volumetric measurements. For everything macro-scale, anyways - if a recipe asks for a teaspoon of vanilla extract, nobody will tell you how many grams that was supposed to be and there aren't that many available sources of variance. Much less the case for e.g. "cups of flour"! Flour in particular is highly compressible[3] and many recipes use highly unrealistic estimates of how many grams there are in a cup of flour[4], when telling you how many cups to use. Fortunately, the aforementioned friend also ran a Flour Measuring Science Party and walked away with a spreadsheet. And it turns out that you can hit 120 grams per cup if you carefully scoop the flour in with a fork, but if you just use the cup measure itself as the scoop you're more likely to end up at 140 grams, and deliberately packing it down can get you to 180. Which is to say: always use mass measurements when available. If a recipe website doesn't either default to mass measurements, or provide a toggle, that's a deeply negative sign about its quality. Relatedly...
Own a kitchen scale. You want one that lets you switch between different units and zero out the current weight. This one is pretty good; the ones that cost $12-15 are probably also pretty good.
Disposable shower caps are a huge improvement over saran wrap, when it comes to operations like "cover the bowl containing the bread dough while proofing". 95% reduction in effort. Something like these[5].
You can "prep" good bread dough in less than ten minutes. I recommend Zvi's transcription of the core recipe from The New Artisan Bread in Five Minutes a Day. I've never tried it with all-purpose flour; I recommend just purchasing bread flour. You may notice that the linked recipe uses volumetric measurements. Having made that recipe probably 20+ times now, and having a sense for how minor variations in flour/water ratios affect the dough, I can now provide you with reliable mass measurements instead, as well as other improvements and notes:
Fast[6] Bread
Follow steps 1 - 3 in Zvi's post. You can speed up step 4 by heating your oven up to ~110F (or using a "warming mat") and tossing the covered dough in for ~75 minutes. You'll likely need to dial this process in yourself, so give it an extra 15 minutes at room temp the first couple times you do it to check how much more it rises - if it's a noticeable amount it probably wasn't quite there.
Zvi's step 5 says "Put it in the refrigerator and use as needed. It should be good for at least two weeks." I would go further and say that there's a noticeable improvement in bread quality after the first 12-24 hours of refrigeration, compared to making it immediately after the first rise. Fermentation will continue (slowly) over the coming days, which many but not all people regard as a positive.
Zvi then goes on to the actual baking instructions. Here are some additional notes of mine:
Salted butter contains meaningful amounts of salt, but this is usually fine if you don't have unsalted butter. Table salt is about 40% sodium, so take the sodium quantity in the salted butter (don't forget to multiply the "per serving" by the number of "servings" in the quantity of butter you'll end up using) and multiply it by 2.5 to see how much you should reduce the quantity of "salt" (by mass) you add. A teaspoon of table salt is roughly 6 grams, so if a recipe calls for a stick of unsalted butter and a teaspoon of salt, and you only have a stick of salted[7] butter, just reduce it to two-thirds of a teaspoon[8]. Relatedly...
Pay attention to whether the recipe is asking for kosher salt or table/fine salt. Salt is one of those cursed ingredients that even the good recipe websites will generally only provide a volumetric measurement for, generally in teaspoons. The usual conversion ratio is to cut the volume of salt in half when going from kosher to table salt - that is, you should use half a teaspoon of table salt to substitute for a full teaspoon of kosher salt[9]. But, also, many recipes are not that sensitive to the exact amount of salt and if you overdo it by 20-30% you probably won't notice. (You might if you underdo it.)
Cakes are often a bad trade-off in terms of effort vs. reward. They often take dramatically more time than other things that people tend to like about as much, so if you're making a cake you should probably be trying to make something that doesn't have a relatively close substitute in the rest of dessert-space. If you just want "chocolate dessert" make these muffins. Other people might suggest brownies, but, ugh. There are some exceptions: with a bit of iteration and experimentation, Claude and I came up with a surprisingly good (and vegan!) chocolate cake recipe that you can probably prep in less than 30 minutes:
Dark Wacky Cake
INGREDIENTS
STEPS
1. Preheat oven: Preheat oven to 350°F (175°C). Lightly grease a 9x9 inch pan or line with parchment paper.
2. Combine dry ingredients: In a medium bowl, whisk together all of the flour, sugar, cocoa powder, baking soda, salt, and espresso powder. Sift if the cocoa is lumpy—Double Dark tends to clump. Make sure the baking soda is evenly distributed throughout.
3. Combine wet ingredients: In a separate bowl or large measuring cup, combine cold water or cold coffee, olive oil, vinegar, and vanilla extract. Stir briefly to combine.
4. Mix batter: Pour the wet ingredients into the dry ingredients. Stir until just combined—you'll see some fizzing as the vinegar reacts with the baking soda. Don't overmix; a few small lumps are fine. The batter will be thinner than a typical cake batter.
5. Bake immediately: Pour the batter into your prepared pan right away—the leavening reaction is happening now. Bake for 28-32 minutes, until a toothpick inserted in the center comes out with moist crumbs (not wet batter, not bone dry).
6. Cool: Let the cake cool in the pan for 10 minutes, then serve directly from the pan or turn out onto a rack. Top with powdered sugar, ganache, or eat plain.
Relatedly, modern frontier LLMs[10] are surprisingly useful cooking assistants:
Recipe websites are deeply unreliable for estimates of "prep time". Generally their estimates are dramatic underestimates for total time from "getting off the couch" to "thing goes in oven", even if you're an experienced home baker. Frontier LLMs also often make mistakes like this, at least with the kind of naive prompting that I've tried so far[12].
Thanks to Drake Thomas for getting me into baking, introducing me to Smitten Kitchen, buying three oven thermometers, and hosting a Flour Measuring Science Party.
$5 - $15 apiece
Though this is less likely with convection turned on.
And therefore high-variance when measured by volume.
120-130 grams is the most common translation that recipes use, for AP flour.
I'm not sure I've ever been the one to buy them, but they're pretty undifferentiated except for size and you probably just want the lowest unit cost.
To prep! Your fastest end-to-end time for actually having bread that you could even in theory eat is something like 2 hours, and that'd be cutting a lot of corners.
Very often 90mg of sodium per 14g serving, or ~720mg per stick (113g).
0.72 * 2.5 = 1.8; 1.8/6 = 0.3
This is apparently only true for the Diamond Crystal brand of kosher salt; you multiply by 0.75 rather than by 0.5 if translating from Morton's kosher salt. But apparently most recipes assume Diamond Crystal, so "cut it in half" is usually correct. The additional facts in this footnote I learned from LLMs while writing this post, so consider taking them with a grain of salt.
Opus 4.6 and ChatGPT (5.4), both with extended thinking enabled, at the time of writing this.
Often this will be down to "tastes vary; if you want more [x] do this, otherwise do [y]".
I haven't tried anything more clever than "Give me recipes following [constraint x] that take less than 45 minutes", or "How much prep time will [recipe] take?"