Difference between revisions of "AI curse or blessing?"
From CitconWiki
Jump to navigationJump to search (Discussion on AI usage, problems, impact on the future) |
|||
| Line 1: | Line 1: | ||
| − | What are the attendants using AI for? | + | === What are the attendants using AI for? === |
| − | + | * Enhanced search engine (when latest data not necessary | |
| − | + | * Assistant / bounce ideas of it to clarify own thinking, and get other viewpoints | |
| − | + | * Specific task code generation, 0 to 60% quickly. (the part after that is the harder bit) | |
| − | + | * Using it as a text assistant, making your text more compelling | |
| − | + | * Implement low/medium -difficult code, by providing API documentation and context | |
| − | + | * Social media content generation | |
| − | + | * Cooking recipies | |
| − | Assumptions on effect of workforce | + | === Assumptions on effect of workforce === |
| − | + | * Heavily reduce need for junior developers | |
| − | + | * Increased projects / investments as assumed higher productivy, lower overall cost, faster to try out ideas, new solution possibilities | |
| − | + | * New job: Prompt engineer | |
| − | + | * Major demand for skillset to train models on company data | |
| − | + | * New skill demand: Detecting fake/AI -generated data, text, images etc. | |
| − | + | === Changes to developer/software professional role === | |
| − | + | * Generating code is one segment of the role | |
| − | + | * Others are likely emphasized in value as AI spreads: communication, debugging, understanding tech, product and market context, creativity, software engineer practices, architecture, logging/observability to speed up AI assisted iteration | |
| − | Other implications | + | === Other implications === |
| − | + | * Accelerate "Dead Internet" (https://en.wikipedia.org/wiki/Dead_Internet_theory) | |
| − | + | * AI vs. AI. Usage of AI to generate actions, and counterpart using AI to counter these actions. Both malicious and everyday. | |
| − | + | * Education: Emphasized need to navigate and validate information validity. | |
| − | + | * EU regulations on the way for AI limitations, US only talk at the moment. Will companies comply in their products by strictest market (EU) or ship regionally configured solutions? | |
| − | + | * "Generated by AI" and "Information is incorrect" are both interesting aspects but they are not the same. | |
| − | + | * Everything homogenised? Knowledge, music, etc. | |
| − | Evolution of assisted writing | + | === Evolution of assisted writing === |
| − | + | * Spell check | Corrects you | | |
| − | + | * Grammarly etc | Rephrases for you | | |
| − | + | * Chat GPT | Generates text with you setting the context | | |
The responsibility remains on the author/committer/approval stamper as before. | The responsibility remains on the author/committer/approval stamper as before. | ||
Latest revision as of 21:54, 27 May 2024
What are the attendants using AI for?
- Enhanced search engine (when latest data not necessary
- Assistant / bounce ideas of it to clarify own thinking, and get other viewpoints
- Specific task code generation, 0 to 60% quickly. (the part after that is the harder bit)
- Using it as a text assistant, making your text more compelling
- Implement low/medium -difficult code, by providing API documentation and context
- Social media content generation
- Cooking recipies
Assumptions on effect of workforce
- Heavily reduce need for junior developers
- Increased projects / investments as assumed higher productivy, lower overall cost, faster to try out ideas, new solution possibilities
- New job: Prompt engineer
- Major demand for skillset to train models on company data
- New skill demand: Detecting fake/AI -generated data, text, images etc.
Changes to developer/software professional role
- Generating code is one segment of the role
- Others are likely emphasized in value as AI spreads: communication, debugging, understanding tech, product and market context, creativity, software engineer practices, architecture, logging/observability to speed up AI assisted iteration
Other implications
- Accelerate "Dead Internet" (https://en.wikipedia.org/wiki/Dead_Internet_theory)
- AI vs. AI. Usage of AI to generate actions, and counterpart using AI to counter these actions. Both malicious and everyday.
- Education: Emphasized need to navigate and validate information validity.
- EU regulations on the way for AI limitations, US only talk at the moment. Will companies comply in their products by strictest market (EU) or ship regionally configured solutions?
- "Generated by AI" and "Information is incorrect" are both interesting aspects but they are not the same.
- Everything homogenised? Knowledge, music, etc.
Evolution of assisted writing
- Spell check | Corrects you |
- Grammarly etc | Rephrases for you |
- Chat GPT | Generates text with you setting the context |
The responsibility remains on the author/committer/approval stamper as before.
