Emma Hudson

Top 10 Prompt Engineering Dos and Don’ts Prompt engineering is a brand new discipline for writing and optimizing prompts to efficiently use large language models (LLMs) for many applications and research topics. Prompt engineering aims to structure the prompt text you send to the LLM so that it is correctly interpreted and understood. Well-crafted prompts direct the LLM to produce accurate, relevant, and contextually appropriate responses. These prompt engineering dos and don’ts will help you improve your generative AI work.

2

The latest in OpenAI vs Publishers... Soon, ChatGPT users will see links to FT articles that provide extra context to their queries. It's an interesting development, and it made me wonder: if deals like this become widespread, will writing now need to be optimised for AI discoverability, as it is today for SEO? How do you make sure the chatbot cites or links to your article instead of a competitor's? Story via TechCrunch: https://lnkd.in/eXUrCsdn

6

There's an inverse relationship between the following words/phrases and the quality of a corporate White Paper/blog. In my 20 years' experience in journalism and advising companies on their thought leadership and PR, the higher the frequency of these repeat offenders, the more banal and badly written the text. - Software solution provider - Leverage - Disruptor - Industry 4.0/5.0 … Web 4.2 etc. - Digital agility - Delighted to announce - With technology moving faster than ever/In our digital world - AI won't take your job, but someone with AI skills will - Groundbreaking - Tailored solutions (for software suppliers) What have I missed? My advice: Recruit an editor from a journalism background to cut through your corporate waffle. (Or use a freelancer.) Removing filler words and phrases will make a substantial improvement in the quality of your business writing (easily 20%) and help you stand out from your competitors.

24

15 Comments

Cracking insights here on the broad implications of the recent licensing agreement between the Financial Times and OpenAI. Media companies rightly should get paid for their IP, but the author doubts whether they're getting the value they deserve. Moral: tech giants generally win. Interesting commentary too about the hype around AI - throw 'AI winter' into your preferred search engine and you'll see why this is a thing. Is AI really as transformative, disruptive and world-changing as Big Tech tells us it is? We've heard these words before about lots of different technologies. Or is it just another way to concentrate more tech money in a handful of big firms? There's a narrative developing about AI's ability to enhance productivity at scale and real concerns about its environmental impact - see the comments for a recent AFR piece on the expected increase in electricity demand. Talking of energy, I'm bed-ridden with flu/Covid today so considering all this through the prism of a fantastic free AI Ethics course from the University of Helsinki. See the comments for a link - highly recommended! #AI #Ethics #Technology #PublicPolicy #Governance #Environment #Media

1

2 Comments

A long tail... or a sideways skyscraper? It's perfectly possible to have a great career in 'marketing'/'advertising'/'media' without ever being featured in a trade publication or winning an industry award. The internet has completely changed the game; while big brands, media owners, and agencies are important, their collective slice of the pie is being outgrown. This is a great thing — a sector that is being oxygenated from below instead of dictated to from above (leaving aside the well overdue issues of regulating big tech platforms that enable new creators to do their business...) But it presents challenges for a niche business journalist like me: we are trained to cover 'big stories', feature 'big companies', and talk to 'big names'. It's based on an audience-centric model, in which we try to aggregate what is interesting to widest possible number of people. As I write in my The Media Leader UK column this week, I suspect the long-term approach for us will be related to Cannes Lions' decision this year to introduce a Creator Pass but on a much wider scale i.e. we work harder to create a range of networks with lots of communities within the long-tail. This is something I've been openly thinking about for a while. How can trade media best harness the growing generation of micro-creators, influencers, consultants, strategists and, writers? As ever, please contact me privately (DM or [email protected]) or publicly by leaving a comment below. https://lnkd.in/e9wjrwvj

11

1 Comment

This is an excellent, subtle article on AI transparency. Simply labelling AI content, as some politicians are advocating, is not a comprehensive solution to issues around trust and misinformation. I highly recommend reading it, but some points that really hit home with me were: - Current labels based on a technical description (AI was used to make this) don't reflect audiences' actual concerns (do those changes matter to me). - "AI labels may drive skepticism about authentic media, diminish trust in truthful content, and also fail to provide utility to communities across cultures and contexts." - Not enough research has been done into the impact (good and bad) of labels, and the final form of labelling should not be left up to a few major platforms, but instead but a collaboration with researchers and civil society groups. I've seen the same issue in misinformation literature on labelling inaccurate content. Some interventions (like asking someone to think twice before sharing) work well, but others might backfire if they are perceived as biased or censorship. https://lnkd.in/gh5dmYEJ

6