Community

I think many of the banks have their heads in the sand. Barcleys for example "If you find it difficult to remember password information, we can confirm your identity using voice recognition technology when you call us, so you don’t have to go through security questions". It's pretty easy to get past this step backward in security with today's voice-cloning technology. The voice can be cloned from a short recording, a family video for exmaple and voila, you know the family security questions, you have the voice and you're in.
Frankly the only way we're really going to protect from most of this is multi-factor and digital signing, things we've already used and now trying to replace with easier options. Easier means less secure and more vulnerable, security is NOT easy.

This could well be a game-changer in the payments world, very exciting.

The perfect storm, Anne Boden, AI and FinTech.

As I replied on X, this is an excellent decision. I spent many years of my life at Paribas and then BNP Paribas leading FX trading and then global architecture. It's critical that EU AI like Mistral gets support from major institutions like BNPP. It's equally critical that banks make early moves to make use of AI and I think the innovation of Mistral will play a major role in moving BNP Paribas into the lead in AI.

It's good to see the larger banks taking the reality of AI seriously, they lagged behind with the internet, mobile etc. With the data the larger banks sit on they could actually take advantage and move ahead of some of the FinTechs.

I have to say, I find this rather funny. I wrote a little program for simulating activity back in the late 80s / early 90s. You could effectively sleep at the keyboard and if anyone came over you could just hit any key and it would write code (my job at the time). I could do my week's work in half a day so there was little else to do. This sort of thing is entirely down to the employer, if they're employing people who are obviously not needed and no one notices their lack of productive activity then who's to blame? People will always find ways to be lazy and to get around amateur attempts to monitor activity. Again, surely the measure of productivity is results? This reflects on the employer, not the employees.

This brings the simplicity of Open Banking to the US but unlike true Open Banking, Apple takes a healthy cut and the banks are disintermediated. Frankly the banks are to blame here, Apple are slow but the bank are slower.

Most of the rest of the world have an opportunity with Open Banking but the banks continue to fight it. They too will lose this battle unless they start moving quicker and accepting that technology is going to change the world of payments.

I'm a huge fan of Mistral for many reasons; It's European to start with and therefore not (potentially) restricted by the new ridiculous laws being proposed in California. It is by design, multilingual which is critical for a European market. it is a serious contender not just in the generalised GPT race but very specifically in open source LLMs. Mistral's open LLM's are not just amongst the best, but for many months have been undisputedly THE best. The Mistral 7B (v0.3), release less than a week ago, is incredible for function calling, beating Meta's Llama 3 released only weeks earlier.

Je me sens même obligé d'écrire en français, je suis très fier d'avoir une société comme Mistral en Europe.

I believe Mistral will play a strong part in the future of privacy in European, and global, AI. Viva Le Chat!

It's a lovely idea but simply not feasible or even technically possible. It's like putting in back-doors for encryption, it's just not mathematically possible without fundamentally breaking cryptography. Firstly the LLMs or GPTs being referenced are global, so who's regulations to you build into the model? Secondly LLMs don't follow rules like that, they can be tuned towards a direction but like people they find loopholes and unless the rule is rock solid and unambiguous, which they rarely are, they simply won't work on LLMs. Censored LLMs, which is effectively what this is suggesting work badly, you're training it and then un-training it. We saw what happened with some of the recently censored models proposing black German soldiers in 1943, Native Indian founding fathers of the USA, all in the admirable effort of inclusion but failing. What is needed is better management of LLMs, UK and EU should be using privately hosted LLMs and then frameworks around those to assert compliance and adherence to regulatory practices. This is a hybrid of LLMs, RAG and traditional integration. It can be done but not in the way suggested here.

Welcome Saudi Arabia, at the speed the UK banks are moving with this you will be well ahead by the end of 2023. Open Banking is the future of payments and it's great you've joined the club.