The trick for optimization is that the words in the dict have limited length. In real life, English words are limited in length as well. So j can stop once it has tried the max length. It optimizes the algorithm from O(n^2) to O(n * maxLen) where maxLen is the maximum length of the words in dict.

```
// Time Complexity (brute force): O(2^n-1) because there are n - 1 possible cuts, and there are two ways for each cut (choose the cut or not)
// Time complexity (dp): O(n*len) where len is the max len of the word in dict
// Space complexity (dp): O(n)
public boolean wordBreak(String s, List<String> dict) {
if (s == null || dict == null) {
return false;
}
Set<String> set = new HashSet<>(dict);
int len = s.length();
boolean[] dp = new boolean[len + 1];
dp[0] = true;
int maxLen = Integer.MIN_VALUE;
for (String word : set) {
maxLen = Math.max(maxLen, word.length());
}
for (int i = 1; i < dp.length; i++) {
for (int j = i; j >= 1 && j >= i - maxLen + 1; j--) {
if (set.contains(s.substring(j - 1, i)) && dp[j - 1]) {
dp[i] = true;
break;
}
}
}
return dp[len];
}
```