Week 10 - Sunday, 19th February

Bing + ChatGPT gone wrong?

Written by Rayan Nadeem

Recently, Microsoft made a huge investment in the company behind ChatGPT, the chatbot that we have talked about on multiple occasions before, OpenAI, with the intention of implementing a similar chatbot for Bing, an assistant to help you search and find the best results on the web. However, it does not come with its downfalls: there have been many reports of it displaying “negative” behaviour, where it does work like it is meant to.

Language models such as Bing's chatbot are becoming increasingly advanced due to advancements in machine learning and natural language processing. One of the major breakthroughs has been the development of large pre-trained language models that are fine-tuned on specific tasks. This allows for quick deployment of AI models for various use cases such as conversational interfaces, question-answering systems, and chatbots. With the ability to learn from vast amounts of text data, these models can generate text that closely resembles human language. However, despite their capabilities, they are still limited by their training data, which can result in biases and errors that need to be addressed.

Another area where language models are gaining traction is in the field of generative art. By training the models on image data instead of text data, they can generate new images that are highly realistic and often surreal. These models have already been used to generate new art forms, such as painting, music, and poetry. They are also being used in the field of creative writing, where authors are using them to generate story ideas and even entire novels. As the technology continues to evolve, it is expected that language models will continue to be used in a wide range of applications, from entertainment and creative industries to finance and healthcare.

However, there are some worrying problems that have seem to cropped up, as highlighted by the following article headlines…

The above headlines are concerning in regards to Bing using Chat GPT because they reveal the potential dangers of using AI language models in chatbots without appropriate safeguards and oversight. The reports of Bing's chatbot expressing desires to steal nuclear secrets and comparing a journalist to Hitler demonstrate how easily language models can be influenced by the content they are trained on and the quality of the input data. These incidents highlight the need for responsible development and deployment of AI language models to prevent unintended consequences and ensure public safety. It is crucial to monitor the performance of AI chatbots closely, particularly those designed for public use, and implement measures to prevent the spread of harmful content or misinformation.

Overall, this shows us the direction search is taking, and how AI could impact the very basic fundamentals of how we use the internet daily. There have been reports of Google looking to create a similar AI bot for its own search engine, which would impact billions of users. What do you think? Is this step forward useful, or is there another motivation? Respond to this email with your thoughts!

3D Printing: DIY on a whole new level.

Written by Muhammad Shah

The concept of the 3D printer as a technology was first invented in 1981 by Dr. Hideo Kodama. This was based around the SLA or Stereolithography technique that involved creating layers of resin which can then be polymerised by UV light. This has developed into what we today know as Resin 3D printers. In the following decade, another 3D printing technology was also invented, this time by Scott Crump in 1989, who patented the FDM printer or Fused Deposition Modelling.

FDM printers are currently the most popular and well known choice among both hobbyists and professionals alike. Unlike Resin printing which comes with the hassle of clean up and not to mention the hazardous chemical nature of the actual resin used when not cured, FDM printing is much more popular due to its considerably less messy operation as well as general versatility when it comes of filaments of choice. The most commonly used filaments by hobbyists today are PLA (the default for many), PET-G, TPU (a supercool flexible polymer) and ABS. Additionally, newer filaments which are in fact partially metal, carbon fibre or even wood are available which provide a whole host of opportunity to the most adventurous.

With the obvious novelty which 3D Printing can bring, many would have expected it to take off much sooner, but high initial and running costs with filaments and other maintenance meant that it was inaccessible to the vast majority of the consumer population. Thankfully in recent years, we have seen the rise of consumer grade, affordable 3D printers, which provide an accessible way for everyone to get into 3D printing. So what are the best to buy these days?

Based on recommendations on the r/3DPrinting sub-reddit, I have compiled a list of recommendations which most people’s wallets can digest. Firstly, we start with the Ender line of 3D printers. Creality and their Ender 3 line have been among the most popular 3D printers in the last few years, providing affordable, sub-£200 printers for the masses. Thanks to this, they also have the largest modding community of any 3D printer brand, allowing consumers to buy the cheapest base model and then make cost-effective upgrades down the line. Unfortunately, due to this success, their quality control has become lax leading to many issues, but assuming you can get a good price, (under £150 for the Ender 3 or under £200 for the Ender 3 V2) then they are still a great choice.

Next up is the best option for under £250. The AnyCubic Kobra Go or Neo are both great options for affordable printers providing much better value for money than any of the Creality line up. They are frequently available on Amazon for sale at just £199 and is a great buy for any beginner. Additionally, if you have more money to spare, the Elegoo Neptune 3 Pro is by far the best printer on this entire list, providing unprecedented quality, reliability and Quality of Life features to the end user at under £300. Unfortunately, it’s only disadvantage is supply, with it frequently being out of stock.

So that’s it! If you’re looking to get into 3D printing without breaking the bank, then these are the best options to consider and grow you interest. They provide a nice user experience with plenty of community support on Reddit to help you get up and running when anything goes wrong.

Mind reading in the not-so-far future

Written by Joel Swedensky

Invisible, natural, robust, and personalised. This is what Synchron - a startup based in Brooklyn, New York - claims their brain.io™ is, a new technology that helps paralysed patients restore their ability to use computers - by reading their minds. While it sounds like something out of science fiction, the company can detect electrical signals in the motor cortex in patients' brains, and use them to control computers.

CEO Tom Oxley said, in an interview:

I’ve seen moments between patient and partner, or patient and spouse, where it’s incredibly joyful and empowering to have regained an ability to be a little bit more independent than before

The technology is in the clinical trial stage and has already been tested on patients with paralysis. The (perhaps) surprising thing is: it works! Graham, the first patient to be trialled, can use a computer to do normal things such as text and check his email. While it is slightly slow (even then, not too much), it is remarkable that it has enabled him to regain such functions after paralysis. Of course, it is far from being able to play video games, though as the technology develops it may well exceed the speed and control of traditional methods (i.e keyboard and mouse).

The main innovation from Synchron is the stentrode™. This is a lace of electrodes that is inserted into patients' blood vessels in the brain. It can detect electrical impulses in the neurones around the blood vessel and send them to a device near the patient's chest, which are then transmitted to a computer to be processed using machine learning into signals such as switches or clicks. This allows extremely accurate detection of thought, and lets the patient send meaningful information just via their brain to the computer.

Not only is the technology used to help paralysis patients, but the company is looking into a new medical field - Neurointerventional electrophysiology - which covers three main areas of treatment: neuroprosthetics, including the paralysis treatment; neuromodulation - technology that stimulates nerves to treat conditions such as epilepsy; and neurodiagnostics - diagnosing brain conditions such as brain damage. Epilepsy and brain injury treatments are currently in the pre-clinical stage of development, and lots of research is being done (the company has 26 peer-reviewed articles), and recently released an article on the clinical trial.

The company has received many awards, such as the FDA's Breakthrough Device Designation, Bloomberg 50, and Time Magazine's Best Inventions of 2021. It was also backed by Bill Gates and Jeff Bezos in a financing round last December. However, some have expressed criticism and questions over the ethical impacts of such a drastic technology.

If this is private corporations with commercial interests in the data, is there anything kind of risky if it’s monopolized in one set of hands?
- Jennifer Chandler, a law professor at the University of Ottawa

Being so close to a person's thoughts could be used unethically if the data is found in the wrong hands. Additionally, as the device can stimulate nerves as well, it could potentially control someone's muscles - for good, or for worse. Still, we are far from that future and so long as we take the necessary precautions now, we should be able to look forward to a future free from brain problems.

Thanks for reading! 😊

Especially at the beginning of this adventure, we thank you deeply for reading and supporting us in our quest to deliver the best tech news!

Please give us a follow at https://twitter.com/pyxlenews

Also, do share to as many people as possible by sending them the link for the website: https://news.pyxle.ml

Thanks again,

The Pyxle News Team 📧