Chapter 7 — Key Takeaways

One page. The reference card for every carbohydrate argument you'll ever have.


Retire This

"Simple versus complex." It classifies by molecular chain length when what matters is structure. White bread and lentils are both "complex" and behave nothing alike; an apple is "simple" and behaves better than both.


Use This Instead

The question is never "is this a carb?" It's "how much structure is still around it?"

Intact Refined
Examples Whole grains, legumes, whole fruit, potato with skin, steel-cut oats White flour, white rice, juice, most cereal, syrups
Digestion Slower — enzymes must get past structure Fast — structure already removed
Fiber, micronutrients Present Largely stripped
Satiety per calorie Higher Lower
Energy density Lower Higher
How fast you can eat it Slowly Very fast

300 kcal of apples = 3 apples, 15–20 min of chewing. 300 kcal of juice = 600 ml, 2 minutes. Same category, same energy, completely different outcome.

And it's a continuum, not a binary. Oat groats → steel-cut → rolled → instant: nothing subtracted at any step, glucose response changes anyway, purely from particle size.


Glycemic Index — a good tool, a poor system

Five problems: measured alone and fasted (nobody eats that way) · values unstable across variety, ripeness, cooking, and laboratory · individual variation is large · low GI ≠ healthy (ice cream beats potato) · outcome evidence is modest.

Glycemic load (GI × g carbohydrate ÷ 100) is the better number, because portion matters.

If you're looking up a GI value before eating, the intact/refined question would have gotten you there faster.


How Much Do You Need? Three layers

Layer Figure What it means
Strict biochemical requirement ≈ zero Gluconeogenesis + ketones cover it — but the cost is paid in protein (Ch 6)
RDA 130 g/day The amount that removes the need to manufacture glucose. Not a survival threshold.
AMDR 45–65% of energy A population range, deliberately wide. Not a prescription.

There is no optimum. Populations have thrived from ~5% to over 80% of energy as carbohydrate. Anyone giving you a single number is giving you a preference.


Blood Sugar "Spikes"

In a healthy person, a post-meal rise is physiology, not pathology — it's what insulin is for. Typical: fasting 70–99 mg/dL, peak below ~140, back toward baseline in 2–3 hours.

What matters is failure to clear, not the excursion: rising fasting glucose, creeping HbA1c (prediabetes 5.7–6.4%), prolonged elevations. That's insulin resistance, and it's Chapter 26.


The History, in Two Lines

The low-fat era failed because of what replaced the fat — sugar and refined starch. The low-carb era succeeds, where it does, because of what gets replaced when carbohydrate goes — sugary drinks, refined snacks, ultra-processed food.

Neither era was ever about a macronutrient. Both were about substitution — Chapter 2, wearing a hat.


What to Actually Do

  1. Prioritize intact over refined. This is the whole chapter.
  2. Drink fewer calories. Sugar-sweetened beverages are the strongest negative evidence in the category. If you change one thing, change this.
  3. Eat legumes. Cheap, filling, high-fiber, high-protein, and most people eat almost none.
  4. Don't fear fruit. Whole fruit comes with structure. The fructose argument applies to liquid and refined sources.
  5. Match intake to activity. Neither high nor low is a moral position.
  6. Stop optimizing GI and stop chasing a flat curve unless you have diabetes or prediabetes.

Reading the Label (§7.12)

  • Read Added Sugars, not Total Sugars. Milk and fruit show sugar without anything being added.
  • Fiber ≥ 3 g per serving on a grain product suggests genuinely closer to intact.
  • First ingredient: "whole wheat flour" first means something; "enriched wheat flour" first doesn't.
  • Ignore "net carbs." No legal definition, inconsistent method, and maltitol gets subtracted in full despite being absorbed.

"Whole grain" tells you a subtraction didn't happen. It says nothing about grinding or addition — which is where most of the difference lives.


Common Mistakes (and the fix)

Mistake Fix
"Rice digests to glucose so it's basically sugar" Chemically similar, nutritionally different — structure, fiber, satiety, and what it's eaten with
Organizing meals around GI Use intact vs. refined; it's simpler and captures most of it
Avoiding whole fruit over fructose Whole fruit has structure. Liquid and refined sources are the issue.
Assuming "whole grain" = minimally processed It's a claim about subtraction only
Cutting a cheap cultural staple with no clinical indication Do the ledger — money, time, and whether the household will eat it
Fearing normal post-meal glucose rises That's insulin working
Believing the resistant starch "hack" is large Real, reproducible, a few percent

Verdict Summary

Claim Verdict Why
GI tells you which carbs are healthy 🟡 Unclear / it depends Real clinical tool; poor everyday system — unstable, isolated measurement, low GI ≠ healthy
Blood sugar spikes harm metabolically healthy people 🟡 Unclear / it depends Matters in diabetes; no outcome evidence for flattening curves in the healthy
Carbohydrates make you fat 🟠 Probably false Controlled trials find little difference; the real effect is refined form and eating rate
Whole grains are no better than white bread (phytates, lectins) Not supported Real chemistry, addressed by cooking; outcome evidence goes firmly the other way
Erythritol causes heart attacks and strokes 🟡 Unclear / it depends Real study, high-risk cohort, heavy confounding, endogenous production. Genuinely unsettled.

⚠️ When to See a Professional

Symptoms of high blood glucose — excessive thirst, frequent urination, unexplained weight loss, blurred vision, recurrent infections, slow healing — warrant a physician, not a monitor. Diagnosis uses HbA1c, fasting glucose, or an OGTT, not a consumer CGM, and a consumer device should never be used to self-diagnose or to adjust diabetes medication.


One Thing to Remember

The category "carbohydrate" contains lentils and Coca-Cola. Any claim about the category has told you nothing. Ask about the structure instead.