AI In Software Development Flashcards
What does LLM stand ford?
Large language models
Examples of LLMs being used in software engineering today
GitHub copilot is a code generation tool that uses LLMs to generated code automatically.
LaMDA is an LLM from Google AI that can be used to help developers gather requirements form users
Cohere is a software documentation tool that uses LLMs to generate software documentation automatically.
Benefits of using LLMS for coding?
By automating simple (or even complex) tasks thereby allowing people to focus on the most import tasks
By automatically reviewing code, checking for bugs, security issues etc..
Giving fresh new ideas for algorithms
Drawbacks of using LLMs for coding?
Cannot use when privacy/confidentially is a concern
Vulnerabilities can be found in AI produced code.
Puts the burden on testing and checking - Need to do more in testing/validation to ensure that the code is good
Training uses publicly accessible code. There are not straightforward criteria for judging quality, so how good is the code you get from AI?
Skills. The people using these tools may not have the skills to understand, debug, nor optimize the code.
Bias. Why did the AI adopt a particular solution when there were multiple possibilities? What if the approach it adopted is not the best for your application?
Intellectual property and liability. Training data is often copyrighted. Several legal scholars have argued that the training itself constitutes fair use, but the output of these models may nevertheless infringe on copyrights or violate license terms in the training set.
What s low code?
This is when an application is developed through a graphical user interface instead of traditional hand-coded computer programming
Benefits of low code?
A wide range of people can contribute to an application’s development, not only those with coding skills
Negatives of low code?
Full control of the program features is only available with hand-coding. Security of the resulting program could also be an issue.
