Simorgh

Hallucination is naturally a concern for anyone looking to depend upon LLM-generated answers.

We’ve been testing LLM responses with a CLI, we’re using it to generate accuracy statistics, which is especially useful when the use-case Q/A is limited.

If ‘confidence’ can be returned to the user, then at least they can have an indication if there is a higher quality-risk with a given response.

As someone who works in tech in the UK, this is very scary to me.

The government must back down and ideally apologise.

This will have subtle effects on the sector irregardless of what they do now.

I think this is an issue of perspective.

I know someone who quite recently died of Covid. There is no longer a pandemic, but the virus is still deadly. Their concern is justified.

Economically, the US is critically dependent on China for facilitating consumer demand. The US cannot afford an adversary in China.

Ethically, the US is not offering a compelling framework or alternative that alleviates the issues you raise. This is a real concern.

China has lifted hundreds of millions of people directly out of extreme poverty. This is their prerogative and modus operandi.

The only way forward is cooperation, adjustment and coordination.

I think it’s part of a wider trend in tech towards shipping features without due regard to the efficacy / performance of the product.

Yes, that makes sense.

A comparison could be drawn from the distinction between the Blackberry keyboard and the iPhone touch surface.

Apple’s design approach could well be aiming to make the Vision Pro feel as natural as possible.

In the UK, I’m seeing blue and grey coloured check-marks - the latter is for official accounts.

I think it looks odd to see one account labelled with two checks, which is how an official account can be labelled.

I wonder if the strategy of multiple checkmarks runs the risk of ‘breaking the UI’. Currently the official check-marks appear to be incorrectly aligned to the ‘Official’ label.

I wonder if so many colours would simply look overwhelming and confusing.

I’ve been experimenting with voice-interfaces where typing is replaced by talking, but I find it hard to transition users to voice - we ‘seem’ to prefer typing to talking.

I wonder if this will change.

I personally struggle with LeetCode-style coding tests.

It would be interesting to hear other people's experiences. I'm a Front-End Engineer with experience in Machine Learning. I've built Deep Learning Neural Networks and am proficient with NextJS and ReactJS, in particular.

I'm working on a voice computing project to bring text-to-speech to new audiences.

I think it's totally under-utilised and an easy-to-use tool of generating voicegrams could be super-sharable and - more importantly - useful from a cognitive perspective.

I think that cognitive differences could be levelled using text-to-speech! And provide an alternative and measurable way of consuming quality content.

My email is in my profile!

I've been playing around with the IBM TTS for audiozing research papers. I think that research should be more accessible and I think that the audioform is the way forward.

The comprehension of powerful ideas should be easy, effortless and intuitive. Listening to content has the double benefit of ensuring that talking about complex subjects is all the more familiar. If you've been introduced to a topic through speech, talking about it is all the more natural.

Hi. I'm the founder of Winterdelta. If you want to request the transcription (audioization) of any particular paper that you are reading, then message me through the site!

I quite enjoyed this book because it gave insight into the hard work and dedication that characterised many individuals, such as Max Levchin, the co-founder of Paypal.

As an aside, I also recommend Hatching Twitter by Nick Bilton. This book details the tumultuous roller-coaster ride that was the early days of Twitter. I feel it is an essential read to truly understand the mentality, minds and drive of many within the start-up world.

Protonmail provides my only non-work email address for everything. Is it a good idea to have multiple email accounts with multiple service providers? Seems inconvenient. But then, perhaps I am taking a risk with my current set up.

I love the loading UI: background and the spinner.

Location: Manchester, United Kingdom (willing to relocate, globally).

Remote: No.

Technologies: Python, R, Javascript (Angular, React), Front End.

Focus: Machine Learning predominantly; Torch7, Tensorflow.

email: t.siddiqui [att] warwick.ac.uk.

This book is staggering. I'm on page 1. I've read a couple of dozen lines. This has explained more about calculus then a decade and a half of seeing related concepts.

It's that [-] button next to the username above each comment. It collapses all the cascading comments. It is enabled via JS. If you disable JS, then it disappears.

As mentioned by Alan right at the beginning of his answer, the true visionary behind our current technological age was Licklider. He was an ARPA researcher who, it is argued [1], initially dreamt up the connected network we now know as the internet. Interestingly, when an early iteration of the 'net was put to UCLA academics, they refused to get involved!

[1] Sharon Weinberger, The Imagineers of War (2017).

Ant looks really good as well!

It has a nicely designed DatePicker [1]. Semantic UI doesn't currently have an official implementation of a Date Picker (although there are unofficial versions on GitHub).

[1] https://ant.design/components/date-picker/

no one else seemed to point this out, but your name is really clever :-). 'Real Artists' for the company name and the product 'Ship', as in 'Real Artists Ship', makes me think back to not too long ago when Mr Jobs was still ...

This is really stunning. I can't wait to commence to course. I finished a Masters from a top-50 worldwide university, and frankly, the approach to data science was mediocre at best. The NLP module notes were plagiarised from Stanford and we were quite happy with this! It gave us a break from 20 year old textbooks that set the plodding pace for the Data Mining module. And don't get me started on my deep learning dissertation. The only expert in the uni on the topic got poached by facebook halfway through the project. The universities are finding it difficult to keep up and are resorting to 'interesting' techniques to retain talent - witness the Turing Institute in the UK. They gave out titles to many professors in several universities a year or so ago... as I gather, as a precursor to pivotal data science research.

What can be done to accelerate distribution of hi tech to developing countries?

I too was frustrated by a recent IBM presentation at my uni - there was no technical content.

However I don't think an AI winter is coming, as the author's asserts (towards the end). Applied AI research is driven by orgs like DeepMind and OpenAI, as well as academics. This powerful symbiosis of industry and academia will, in my opinion, continue to yield technological breakthroughs in the coming years and decades.

I'm using this these days. I find it incredibly flexible.

I thought the most interesting news from ATAP (last year) was project Jacquard [1]. It seemed very exciting and promising, but it doesn't seem that the tech has made it to market.

[1] https://www.google.com/atap/project-jacquard/

I wondered what your thoughts were on the apple / fbi situation ? How would you expect startups to act if they found themselves in a similar predicament?

It's a bit unfair for the writer to describe a nation's ruling elite as a 'criminal organisation'. You could levy that accusation at every ruling elite, of every country, of every era.

Nonetheless, the House of Saud funds 'schools' throughout Pakistan. These schools are known for spreading an extremist and violent philosophy [1], which clashes with the democratic and enlightened spirit of the educated-elite, at least in Pakistan [2].

[1] https://www.washingtonpost.com/world/pakistan-is-still-tryin...

[2] http://www.theguardian.com/commentisfree/2015/apr/27/murdere...

Facebook made the internet a home for a billion people. Mr Rushkoff is certainly an important voice, but I'll play devil's advocate. We can never know the initial motives or future designs of Zuckerberg et al, nonetheless these actors have built products/services that people demand. They have introduced advanced technology to vast swathes of people, familiarised them with computers, and drawn them into (the periphery of) hacker culture.

Hypothetically, more intelligent/predictive advertising should decrease the quantity of ads we see, and work in the favour of local firms. AI (including predictive-advertising) has the potential for good. People want to support local businesses. With the introduction of localised / native advertising, there is the potential for the business next door to 'get you' as a customer, as opposed to Amazon etc. The 'last mile' problem might work in the favour of small firms.

Hi, I think your resume is fairly strong. You seem to have the right structure.

The thing to remember is that recruiters typically spend about 6 seconds on a resume. And this I hear from recruiters.

The objective section is crucial by the way, and can be improved by succinctly summarising your past. It is often the only thing that will be read.

You should prepare a tailor-made resume for each application, with the name of the job title you are applying for, in each application. This creates a level of personalisation that will put you ahead of other applicants.

If you want more help I'd get the book, 'Great Answers to Tough Interview Questions' by Martin Yate. In part, it describes how to sculpt your resume. Everyone I know who got that book got multiple job offers, and no, I'm not an affiliate for the book!

PG writes that the 20th Century was "a world in which it was socially acceptable to work for Henry Ford, but not to be Henry Ford".

I feel that this viewpoint is still common today. Success is often glamourised, far more than hard work.