daviddoji/pytudes
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Add files via upload

Browse Source
This commit is contained in:
Peter Norvig
2025-12-23 12:45:27 -08:00
committed by GitHub
parent 02168d2451
commit 45562b6690
1 changed files with 2 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
ipynb/Advent-2025-AI.ipynb
View File

@@ -23,6 +23,7 @@
"- *I'm beginning to think I should use an LLM as an assistant for all my coding, not just as an experiment like this.*\n", "- *I'm beginning to think I should use an LLM as an assistant for all my coding, not just as an experiment like this.*\n",
"- *This is a huge improvement over just one year ago, when LLMs could not perform anywhere near this level.*\n", "- *This is a huge improvement over just one year ago, when LLMs could not perform anywhere near this level.*\n",
"- *The three LLMS seemed to be roughly equal in quality.*\n", "- *The three LLMS seemed to be roughly equal in quality.*\n",
"- *I neglected to track the time it took them to produce the code, but it was a lot faster than me–maybe 20 times faster.*\n",
"- *The LLMs knew the things you would want an experienced software engineer to know:*\n", "- *The LLMs knew the things you would want an experienced software engineer to know:*\n",
" - *How to see through the story about elves and christmas trees, etc. and get to the real programming issues*\n", " - *How to see through the story about elves and christmas trees, etc. and get to the real programming issues*\n",
" - *Standard Python syntax, builtin types, and basic modules (e.g. `collections`, `functools`, `typing`, `numpy`)*\n", " - *Standard Python syntax, builtin types, and basic modules (e.g. `collections`, `functools`, `typing`, `numpy`)*\n",
@@ -3955,7 +3956,7 @@
"\n", "\n",
"*The human-written code is about **five times more concise** than the LLM code.*\n", "*The human-written code is about **five times more concise** than the LLM code.*\n",
"\n", "\n",
"*The LOC numbers are total lines of code, including blank lines, comments, doc strings, and the line to make the call to compute the answer. For the human-written code, I count the parsing code against Part 1.*" "*The LOC numbers are total lines of code, including blank lines, comments, and doc strings.*"
] ]
}, },
{ {
@@ -3991,14 +3992,6 @@
" | **TOTAL** | | **1732** | **356** | | \n", " | **TOTAL** | | **1732** | **356** | | \n",
" | **MEAN** | | **75** | **14.5** | |" " | **MEAN** | | **75** | **14.5** | |"
] ]
},
{
"cell_type": "code",
"execution_count": null,
"id": "fdcde1d0-a999-4867-8b08-6e10a75fa385",
"metadata": {},
"outputs": [],
"source": []
} }
], ],
"metadata": { "metadata": {