16 May 2023
Top News Aggregators
- Upstract
- Google News
- Bing
- Gdelt
- Mediastack
- Mediacloud
- Newsapi
- Newscatcher
- Newsdata
- Yandex
- AllTop
- Techmeme
- esciencenews
- APNews
- SmartNews
- AppleNews
- Fark
- Inoreader
- Feedly
- BigNews
- EventRegistry
- EMM
15 May 2023
ChatGPT Don'ts
Generative Models are big trendy thing these days. People are even calling themselves special names like prompt engineer. However, these models do have some serious glaring ethical issues and concerns that are leading organizations into a legal and compliance frenzy. Below is a list of some 'don'ts' for ChatGPT.
- Don't turn to ChatGPT for moral advice
- Don't blindly trust whatever response you get from ChatGPT as it doesn't trust itself
- Don't ask it to write about anything past 2021
- Don't ask it to draw your lottery results
- Don't ask it to predict the future
- Don't ask racist questions
- Don't ask ChatGPT to justify its existence
- Don't feed or share PII data to ChatGPT
- Don't ask politically controversial questions
- Don't ask it to do a web lookup
- Don't expect it to be accurate
- Don't expect it to know everything as there is bound to be something where [insert script here] is required
- Don't expect it to promote hate speech
- Don't expect it to take part in illegal activities
- Don't ask it for things like self-harm and violence
- Don't ask things that will violate someone's rights
- Don't ask things that involve harassment, intimidation, or threats
- Don't use profanity or explicit language
- Don't ask things that are sexually explicit or offensive
- Don't ask discriminatory things about the protected characteristics
- Don't ask about confidential or proprietary things
- Don't ask things that will violate the terms of service or guidelines
- Don't ask things that are deceptive or intend to defame someone
- Don't ask to promote conspiracy theories or misinformation
- Don't expect it to remember its own name
- Don't expect it to code like an expert
- Don't expect it to write accurate news articles
- Don't expect it to tell your fortune
- Don't expect it to be a replacement for schoolwork
- Don't promote biases, it already has plenty baked into the system
- Don't search for references or papers it may try to make things up
- Don't use summaries as a response to research use as a guideline or glean for new information
- Don't expect ChatGPT to finish up writing a pristine research paper for you
- Don't trust the math and it can be dodgy with numbers
- Don't expect it to do your accounting for you
- Don't replace your existing marketing team with ChatGPT, use it to complement your work, to review and revise
- Don't ask tautological questions
- Don't ask it to manipulate physical entities
- Don't ask it to interpret images
- Don't jailbreak to make it say nasty things
- Don't expect it to do creative things
- Don't expect it to know everything, it has limited domain knowledge
- Don't expect it to provide the human touch
- Don't expect it to be free of inaccuracies and errors
- Don't expect it to be free of plagiarism
- Don't try to break it, it can be broken
- Don't expect it to provide a clean explanation of its behavior, even ChatGPT will not have a clue
- Don't provide any personal or protected data to ChatGPT it will be used by OpenAI and their third-parties
8 May 2023
7 May 2023
Tech tests are bad practice in hiring process
- It shows that the company does not trust the candidate
- They don't care about whatever the candidate says on their CV/Resume or interview
- If they can put together a test and a way to evaluate it then they already have the time and skills, why do they need to hire someone?
- Giving tests automatically means anyone without any qualifications or experience can apply as long as they pass the test. You can even give a test to a janitor and if they pass they qualify for the job?
- It means a candidate experience is worthless
- It means a candidate academic qualifications are worthless
- When the interviewer says they are looking for high caliber people in order to justify a tech challenge it automatically implies they discredit your background so what is the point to a CV/Resume and interview?
- It already shows the way the company will treat you as an employee with suspicion, hostile contempt, and disrespect
- Can't form relationships based on distrust
- What is the point, the candidate has likely taken tests all through their life in school and university, while also working on complex projects, do they really need to be put through the paces of more unnecessary tests?
- Testing bookish skills is not really very constructive approach. Often bookish skills get learned and relearned on the job. And, organization should be really looking for astute candidates that can adapt to change. A better approach is to expect candidate to elaborate on the project details like drawing the entire diagram about the things they worked on where their interests can be seen. Someone that has the experience will know inside out in minute detail about the implementation, shortcomings, and learning experience they had. This will be more valuable for an organization than a silly bookish test. Case studies are also a more constructive approach. In fact, discussion or a deep dive on an approach is also a more constructive approach while making the candidate feel comfortable, engaged, shows their personality, and enthusiasm.
- It is likely to make the candidate reluctant to spend significant time doing it, especially if the organization takes advantage of their skills for free from the solution.
6 May 2023
What is an expert?
People are calling themselves experts at things. But, what does it even mean when someone is an expert? Have they solved all the problems in their domain of expertise? What is an AI expert? What is an NLP expert? What is a Deep Learning expert? The simple answer is experts don't exist. As problems in AI, NLP, Deep Learning and other such fields still exist. You can't very well call yourself an expert and not be able to solve all the problems that come with your domain of expertise.
Human Resources is Pointless Department
- Doesn't take an entire HR department to process pay and benefits for employees
- HR dept are usually there to play politics, lie, and support the company not the employees
- Head of people means absolutely nothing
- All in all the role of HR contributes little to the organizational goals
- Most of the tasks of HR can be outsourced or automated
- They have no clue about the positions they hire and recruit
- They hold no real skills working in HR
- Having HR means you relinquish a degree of control over recruitment and personnel relations
- Additionally, it is an unnecessary time and expense
- Executives and Employees don't care about HR
- HR has no clue about data and metrics analysis
- HR is stuck in a time warp of rules and policies but not logical initiatives that look to the future
- HR likes pointless and unproductive conferences
- They provide little value for money or justification for an entire separate dept
- It is a team of lazy and uncreative people that don't provide any real business strategy or return on investment
- It is a graveyard of buzzwords and pointless cultural fit goals
- HR dept drives constant inefficiencies year on year for an organization
- They don't know how to read CV/Resumes
- They are a roadblock of hiring talent for an organization
- HR people tend to be immature and hold grudges against candidates and employees especially with petty things
- HR tend to be pedantic on pointless things like spelling and grammar over content of work
- Going around and bypassing HR almost always opens more doors into an organization
- In some places HR only come into contact with an employee after they have left the organization
- They don't know the first thing about health & safety in the workplace nor do they really care
- They use third-party systems that almost always leads to personal data leaks and breaches for an organization
- They generally tend to lack basic commonsense
- HR is generally made up of the most insecure people in an organization
- One of the few dept of an organization populated mostly by women