Wow!! It’s such an amazing tool – beating back ChatGPT hands down – that I had to try it out on some conundrums I faced in the past. Note that on opening this post, a video starts playing automatically at 11:11 min and stops automatically about 6 mins later. To not have the video start automatically, simply do not visit this article. You have choice in doing that. Viewers can opt to stop the video by clicking on it, if they access this article.
How does DeepSeek think?
How does a plane fly in the sky? How does a microwave oven work? You could spend your life trying to understand those sorts of things if you wish, before taking advantage of them. You have choice!! [I nauseatingly repeat this at every turn in recent times]. Ooooo.. The video below has the word ‘hacked’ and ‘banned’ in it. OMG!! OMG!! Terrible – bite fingernails and stay away. Say to yourself, “That’s dangerous. I don’t want anything to do with hacking! Switch that off! How dare you!” Do I give a flying flamingo https://youtu.be/1KI9Ruv-BDI – I do not!. You were told not to visit this link if you did not want the video to start automatically!
Please accept YouTube cookies to play this video. By accepting you will be accessing content from YouTube, a service provided by an external third party.
If you accept this notice, your choice will be saved and the page will refresh.
The issue I put to DeepSeek was as follows:
Let’s say I’m setting up a conference for 50 people in the UK. I want to find the best location for the conference, prioritising the minimisation of travelling time. The means of transport could be quite variable. Some people may use cars (including taxis), some may use the bus if they live quite nearby, those living further afield may use train or air transport. You can imagine that some people living quite far away but travelling by air would actually make the journey in less time than someone travelling by road (due to traffic etc). Of course, I won’t know what means of transport individuals may choose. So that led me to think about finding a ‘centre of gravity’ based on distance only based on postcodes. But I don’t know if that is the best method. Give me your analysis.
DeepSeek thinking
This was how DeepSeek thought about it in just under 3 minutes of thinking time.
It’s not necessary to read every line of the following. The important points are: reasoning, evaluation, understanding limitations and practical constraints, coming up with a plan and suggesting alternatives. You’d have to pay some ‘fool’ hundreds of pounds (sterling) to come up with that and it would probably be written in incoherent language. 😖
To determine the best location for your conference that minimizes overall travel time for 50 attendees, you need to balance several factors, including the geographical distribution of attendees, their likely modes of transportation, and the accessibility of the proposed location. Here’s a step-by-step analysis of your problem and potential solutions:
1. Understand the Data
- Postcodes: Collect the postcodes of all 50 attendees. Postcodes in the UK are precise and can be converted into geographical coordinates (latitude and longitude) using tools like the UK Postcode API or open datasets.
- Travel Modes: While you don’t know the exact travel modes attendees will use, you can make reasonable assumptions based on distance:
- Short distances: Likely car, bus, or train.
- Medium distances: Likely train or car.
- Long distances: Likely air or train.
2. Calculate the Geographical Centre
- Centre of Gravity: This is a common method to find a central point based on postcodes. It involves:
- Converting all postcodes to latitude and longitude coordinates.
- Calculating the arithmetic mean of these coordinates to find the geographical midpoint.
- Limitations: The centre of gravity only considers straight-line distances and doesn’t account for travel time, transportation networks, or real-world constraints like traffic or flight availability.
3. Refine the Approach for Travel Time
- Travel Time Matrix: Use tools like Google Maps API, TravelTime API, or OpenRouteService to calculate travel times between each attendee’s location and potential conference venues. These tools can account for different modes of transport and real-time traffic conditions.
- Weighted Travel Time: Assign weights to different modes of transport based on likely usage. For example:
- Short distances: Assume car or public transport.
- Long distances: Assume air or train.
- Optimization Goal: Choose a location that minimizes the total travel time for all attendees.
4. Consider Practical Constraints
- Transport Links: Choose a location with good transport links, such as a city with a major airport, train station, and road network.
- Accessibility: Ensure the venue is accessible for all attendees, including those with disabilities.
- Cost: While not your primary concern, consider the cost of travel for attendees, as this may influence their choice of transport.
5. Tools and Methods
- Geographical Information Systems (GIS): Use GIS software like QGIS or ArcGIS to visualize attendee locations and potential venues.
- Optimization Algorithms: Use algorithms like k-means clustering or p-median optimization to find the best location based on travel time.
- APIs: Leverage APIs like Google Maps, TravelTime, or OpenRouteService to calculate travel times programmatically.
6. Step-by-Step Plan
- Collect Postcodes: Gather the postcodes of all 50 attendees.
- Convert to Coordinates: Use an API or dataset to convert postcodes to latitude and longitude.
- Calculate Centre of Gravity: Find the geographical midpoint as a starting point.
- Identify Candidate Venues: Shortlist potential venues near the centre of gravity.
- Calculate Travel Times: Use an API to calculate travel times from each attendee’s location to each candidate venue.
- Optimize: Choose the venue with the lowest total travel time.
- Validate: Check the venue’s accessibility, capacity, and cost.
7. Example
- Suppose most attendees are in London, Manchester, and Edinburgh. The centre of gravity might suggest a location in the Midlands, such as Birmingham.
- Using a travel time matrix, you might find that Birmingham is optimal because:
- It has a major airport and train station.
- It’s accessible by road for those driving.
- It minimizes total travel time for all attendees.
8. Alternative Approaches
- Median Location: Instead of the centre of gravity, use the median location to reduce the impact of outliers (e.g., attendees in remote areas).
- Clustering: Group attendees by region and choose a venue that serves the largest cluster while minimizing travel for others.
Stupid butted in!
| # | Participants | Responses |
|---|---|---|
| 1 | Stupid: | I don’t understand. I think you’re up to no good! |
| 2 | CW: | Do you even think? |
| 3 | Stupid: | No need to be insulting. |
| 4 | CW: | That’s not an insult. It’s a fact. You are not capable of sensible thought most times. I know because I designed you that way. |
| 5 | Stupid: | Right – you’re trying to say that people are stupid. Isn’t it? |
| 6 | CW: | No. I said that AI is a lot smarter than most people. That’s an evidence-based fact. |
| 7 | Stupid: | What you mean? |
| 8 | CW: | ChatGPT which is an AI passed the American Bar exam, in July 2022. |
| 9 | Stupid: | So it now works in bar – who cares! |
| 10 | CW: | You fool! That’s an important law exam. It aced the exam at a 90th percentile. |
| 11 | Stupid: | So now it’s a lawyer. Big deal. |
| 12 | CW: | Chrysst! Most people don’t pass the Bar exam. Most people don’t study law. |
| 13 | Stupid: | Okay – so it’s not a lawyer. It’s just a play thing for people like you with time to waste! |
| 14 | CW: | How do you know that? |
| 15 | Stupid: | Well – I think so – so I’m right. |
| 16 | CW: | Groaaannn! Why do I let you out on this blog? Oh – I should know. |
| 17 | Stupid: | Did I do a good job? |
| 18 | CW: | Yes you did. Now back in the box! |
Conclusion
While the centre of gravity based on postcodes is a good starting point, it’s essential to refine your analysis by incorporating travel time data and practical constraints. Using APIs and optimization tools will help you find the best location that minimizes overall travel time for your conference attendees.
Reflection
The above and many other explorations with DeepSeek and other AIs has informed me that this thing is no ‘calculator’. It has and understanding of situations once given the appropriate context of a problem. But even if it is not ‘human understanding’ it exceeds the average of the majority of humans’ understanding. How dare I say that?! I dare and I do because there isn’t another human being I can call on who would give me such coherent answers.
Did I say that AI is perfect. Look, for the flipping record I’m shouting it out loud: AI IS NOT PERFECT AND MAKES SOME SILLY ERRORS ON A MINORITY OF OCCASIONS!! Why should anybody look for perfection in AI when they wouldn’t expect such from their fellow human beings? What – are imperfections in AI-thinking less allowable than human-thinking? WTF!!
Am I trying to ‘sell AI’ to anyone? ABSOLUTELY NOT! What’s in it for me to sell something like DeepSeek that is totally free!? FFS! Just like with microwaves, washing machines, TVs, handheld calculators, VOIP (e.g. Skype, Viber, WhatsApp) and so on – some people will catch up after 10 years. Everybody has choice – or so they think.
I think that AI software is a fascinating and very novel additional enhancement of human intelligence. If you admit to yourself and others that you are thick as two planks on A, B, or C topics, now is your time to get smarter. Or do you really want to get smarter at all? You have choice!
