diff --git a/README.md b/README.md
index 220d7cd..a57da02 100644
--- a/README.md
+++ b/README.md
@@ -46,40 +46,28 @@ In place modification of files:
 
 ## Benchmarks
 
-**Simple replacement on ~1.5 gigabytes of JSON**
+The commands below assume GNU sed and a POSIX-compatible shell. `sd -p` writes
+replacements to stdout without modifying the input files, so every run uses the
+same input.
+
+**Simple replacement across JSON files**
 
 ```sh
 hyperfine --warmup 3 --export-markdown out.md \
   'sed -E "s/\"/'"'"'/g" *.json > /dev/null' \
   'sed    "s/\"/'"'"'/g" *.json > /dev/null' \
-  'sd     "\"" "'"'"'"   *.json > /dev/null'
+  'sd -p  "\"" "'"'"'"   *.json > /dev/null'
 ```
 
-| Command | Mean [s] | Min…Max [s] |
-|:---|---:|---:|
-| `sed -E "s/\"/'/g" *.json > /dev/null` | 2.338 ± 0.008 | 2.332…2.358 |
-| `sed    "s/\"/'/g" *.json > /dev/null` | 2.365 ± 0.009 | 2.351…2.378 |
-| `sd     "\"" "'"   *.json > /dev/null` | **0.997 ± 0.006** | 0.987…1.007 |
-
-Result: ~2.35 times faster
-
-**Regex replacement on a ~55M json file**:
+**Regex replacement in a JSON file**:
 
 ```sh
 hyperfine --warmup 3 --export-markdown out.md \
   'sed -E "s:(\w+):\1\1:g"    dump.json > /dev/null' \
   'sed    "s:\(\w\+\):\1\1:g" dump.json > /dev/null' \
-  'sd     "(\w+)" "$1$1"      dump.json > /dev/null'
+  'sd -p  "(\w+)" "\$1\$1"      dump.json > /dev/null'
 ```
 
-| Command | Mean [s] | Min…Max [s] |
-|:---|---:|---:|
-| `sed -E "s:(\w+):\1\1:g"    dump.json > /dev/null` | 11.315 ± 0.215 | 11.102…11.725 |
-| `sed    "s:\(\w\+\):\1\1:g" dump.json > /dev/null` | 11.239 ± 0.208 | 11.057…11.762 |
-| `sd     "(\w+)" "$1$1"      dump.json > /dev/null` | **0.942 ± 0.004** | 0.936…0.951 |
-
-Result: ~11.93 times faster
-
 **Line-by-line vs across mode** (1M lines, ~36MB file):
 
 | Command | Mean [ms] | Relative |