LOGIN
התחברות או הרשמה
להמשך הרשמה ידנית – לחץ על כפתור ההרשמה, להרשמה/כניסה מהירה בעזרת חשבון רשת חברתית – לחץ על הלוגו בכותרת
בכדי להוסיף פוסט בבלוג יש להרשם או להתחבר - ההרשמה/כניסה מתבצעת מכותרת האתר.
I listened to 9/22 SE Radio Episode 582: Learning to Program with LLMs with Dr. Leo Porter and Dr. Daniel Zingaro, co-authors of the book Learn AI-Assisted Python Programming, speak with host Jeremy Jung about teaching programming with the aid of large language models (LLMs). They discuss: Why they aren’t worried about students cheating with LLMs How GitHub Copilot de-emphasizes syntax errors and reduces the need to memorize APIs Why they want students to write manual test cases Possible ethical concerns of relying on commercial tools I recommend the podcast. I started reading the bookhttps://www.manning.com/books/learn-ai-assisted-python-programming and found early on (1.6 The skills we need) the following paragraph: “Other skills may take on even more importance with Copilot than without. … We know a lot about testing code written by humans, because we know where to look for typical problems. We know that humans often make programming errors at the boundaries of values. For example, if we wrote a program to multiply two numbers, it’s likely that we’d get it right with most values but maybe not for when one value is 0. What about code written by AI, where 20 lines of flawless code could hide one line so absurd that we likely wouldn’t expect it there? We don’t have experience with that. We need to test even more carefully than before.” I’ve spent much of my career understanding typical problems due to humans writing code. Now, this can get thrown out. But unfortunately with AI assisted code generation “We[…]
1.10.2023 | 12:35 קרא עוד...
A step-by-step guide (with video) to create a test project for a Selenium+RSpec test script in TestWise IDE.Continue reading on Medium »
Fall has arrived. My co-leader and I kicked off our Building a Second Brain book club. We will cover how we are creating a community of note-takers in five brief meetings. Each of us has unique challenges, but we also share commonalities.
Email overload.
Summarizing is difficult.
There are too many intersections.
It is easier to teach what you learned to someone higher up the chain than to learn from the master. My next step will be to put what I learned from The Art and Business of Online Writing and Snow Leopard: Category Pirates into practice.
Writing Online (X/Twitter) with Data
Write online for faster feedback. We often write in solitude, but there are better ways to do it. Every day, I’m going to write.
There are many things I’ve tried, such as a digital garden, different tools like moving from Notion to Obsidian, and linking everything together.
Here is what really works and has been battle tested with real people. Set up and organize your system first. As a result of my own learning, I have developed the CUE framework.
Curate.
Unpack and repack.
Evaluate and review, regularly.
The point of all this is writing is hard yet we need to learn the fundamentals and notice where the friction is difficult.
Start with talking with people.
Hallways conversations, break rooms, and chats are ways to get feedback.
They will suggest how you should sequence.
Think about topics that are timeless and evergreen when deciding what to share. Slack’s[…]
מבט מערכתי לבודקים – Systems Thinking התכונה או מערכת אותה אנו בודקים - אינה מנותקת משאר העולם, תיקון בנקודה אחת – עשוי להשפיע ולפגוע בנקודות אחרות, ולכן על הבודק להכיר את המערכת והקשרים בין חלקיה. כמו…
"בדוק מוקדם ככול הניתן" – אחת המטרות של הבדיקות הינה לספק כמה שיותר משוב ומידע לגבי איכות המערכת מנקודות מבט שונות (פונקציונליות ולא פונקציונליות כמו עמידה בעומסים ושאר יכולות). לעיתים בודקים נוטים בטעות לחשוב כי הדרך…
