How OpenAI is Completely Transforming Finance Right Now (Exclusive Interview w/ James Dyett)

[@raymota5343]

For a minute I thought it was Sam Altman talking until i looked at the video. Excellent interviewer and interviewee - well done!!!

[@CambrianHQ]

Glad you enjoyed it!

[@alexanderbrown-dg3sy]

Working on models myself. I think it’s actually insane Morgan Stanley has been so aggressive with deployment. All models due to positional encoding methods are trash at context utilization, like multi-hop fact synthesis which is MUST for RL use. I have a feeling their next model will have the same limitations, just more capacity, so at least it will be more reliable.

[@gabrielyao9840]

Great questions and inspiring answers, just an excellent interview!

[@CambrianHQ]

thanks for watching!

[@simonnilsson5356]

Would liked to know a lot more details about Klarna’s implementation. Like, how do they, technically, make the LLM execute tasks? Is it agentic, is it integrating with a traditional software to execute tasks etc.?

[@CambrianHQ]

Great questions. We weren’t able to go that deep here but may be able to in the future. I don’t believe it is an agentic implementation.

[@simonnilsson5356]

@@CambrianHQ also the total save doesn’t seem to be 100% accurate. I have industry insight and it looks like they are hiring 200 Customer Service reps through a BPO in Europe. If you’re interested I can provide more details in DM.

[@CambrianHQ]

who do you think benefits the most from implementing AI? Fintechs like Klarna or incumbents like Morgan Stanely?

[@tma2444]

definitely incumbents. A lot of deadweight to be displaced.

[@dhritayyyyyy]

One of the reasons why AI scaling is different from traditional revolutions (approach from Startups to scale to enterprise) is probably because of its dependency on Data. Which startups essentially wouldn't be able to live up to Scepticism around AI is a good thing, because soon everyone is going to realise that the impact is a compounding and an iterative one. And the ones to lead are only going to be the one's who start to adopt early

[@dawithaileleul4299]

37:40 is the visceral aha moment, which is the X factor. The last time I could think of a product with the same level of aha for me personally was when I called Uber for the first time.

[@CambrianHQ]

The demo-ability of chat GPT is definitely an under appreciated aspect of why Chat GPT took off.

[@Sergeo333]

This was really good! very informative!

[@clifftanch]

I have witnessed a number of banks in developing markets try hard to shut down branches because of the cost, so the idea that ATMs lead to hiring of more tellers sounds false and wishful thinking.

[@CambrianHQ]

it is what the data shows. Developed and developing markets are not the same. Here is the data. www.aei.org/economics/what-atms-bank-tellers-rise-robots-and-jobs/

[@clifftanch]

@@CambrianHQ The specific example I had in mind was South Korea, arguably a Semi-developing country. I spent a 30-year career in financial services, almost all in developing countries but now live in the Silicon Valley. When I was on Wall Street during the dotboom, I realized many of my colleagues knew so little about Tech. But now I think the favor is returned: So few in Tech seem to understand financial services. My first experience with an ATM was in 1978, my first year in college. The AEI data clearly show that was when growth in tellers flattened noticeably. I would argue ATMs had everything to do with that. By 2010, the data show that growth in ATMs had also flattened. My comment would be that despite ATMs and despite slow growth in tellers, banks were still unable to make branches pay for themselves.

[@brandonreed09]

​​@@CambrianHQRead the whole article, smh. The amount of tellers needed per branch went down from 21 to 13. What increased was the number of branches, the increase in tellers was just a by product of that. The branches served as a type of marketing provider for the bank. However the branch numbers are now falling with Mobile Banking reducing the need for bank branches as marketers of bank services. Subsequently the amount of tellers is also now in decline even though the tech is being adopted on everyone's phones.

[@MannyBernabe]

Excellent. Great tips on GenAI enterprise GTM. Thank you.

[@CambrianHQ]

for me the GTM advice was the most interesting. for example, why you DO NOT want to run quotas for certain types of companies at certain stages (and why OpenAI does not) 49:15

[@dan-cj1rr]

were all gonna get fked in the 4ss in the end.

[@enterprising9]

Awesome storytelling! Love this interview guys 🎉

[@CambrianHQ]

Thanks for tuning in!

[@WorkAnywhereGo]

Excellent convo!