large language models
New GPU indexes to provide transparency into AI compute
Silicon Data, a company backed by DRW and Jump Trading, has launched its H100 and A100 indexes, providing transparency into the economics of AI compute.
Wall Street hesitates on synthetic data as AI push gathers steam
Deutsche Bank and JP Morgan have differing opinions on the use of synthetic data to train LLMs.
SNB researchers test LLM-based FX trading strategy
Meta’s Llama 3.1 comes out on top at predicting G10 currency sentiment based on news articles.
Waters Wavelength Ep. 329: LLMs and the dead internet theory
This week, Reb and Tony talk about the internet and the problems compounded by AI.
The great disappearing internet—and what it could mean for your LLM
AI-generated content, bots, disinfo, ads, and censorship are killing the internet. As more of life continues to happen online, we might consider whether we’re building castles atop a rotting foundation.
Northern Trust Asset Management enlists network theory to construct alpha signals
The $1.3 trillion asset manager will publish a paper later this year detailing how measuring network effects helps it better understand persistent drivers in trends.
Generative AI brings testing times for modelers
Flagstar’s lead model validator offers some tips for safely integrating LLMs into risk models.
Waters Wavelength Ep. 326: Connectifi’s Nick Kolba
He joins the podcast to discuss the model context protocol, LLMs and agentic AI.
S&P Global details AI partnerships, LLM advancements
The data provider has partnered with Microsoft and Anthropic to use hyperscaler tech to boost its AI offerings.
The industry is not ready for what’s around the corner
Waters Wrap: As cloud usage and AI capabilities continue to evolve (and costs go up), Anthony believes the fintech industry may face a similar predicament to the one facing journalism today.
Academic warns of systemic risk from AI-powered trading
Strategies generated by LLMs exhibit “very strange, correlated trading behavior”, says Lopez Lira.
The Model Context Protocol brings agents to life—along with risk
Waters Wrap: From chat to infrastructure modernization, Anthropic’s MCP offers a ‘bridge’ to agentic AI, but its early days may prove disillusioning.
Agentic AI comes to Bloomberg Terminal via Anthropic protocol
The data giant’s ubiquitous terminal has been slowly opening up for years, but its latest enhancement represents a forward leap in what CTO Shawn Edwards calls, “the way we should talk to the world.”
Deutsche Bank casts a cautious eye towards agentic AI
“An AI worker is something that is really buildable,” says innovation and AI head
LLMs are making alternative datasets ‘fuzzy’
Waters Wrap: While large language models and generative/agentic AI offer an endless amount of opportunity, they are also exposing unforeseen risks and challenges.
Perceive, reason, act: Agentic AI, graph tech used to assess risk
Industry executive Jay Krish is experimenting with large language models to help PMs monitor for risk.
First Citizens used AI to retain SVB customers
The firm’s retention efforts involved using AI to monitor customer behavior and sentiment—including profanities.
Everything’s a chatbot. Soon, your sales trader might be, too
Morgan Stanley, Citi, and Kepler Cheuvreux are among firms considering making their internal AI assistants available to clients.
As agentic AI comes into play, buy side weighs outsourced options
Asset managers wanting to use sophisticated forms of AI must square their interest in the tool with mounting cost pressures.
Deutsche Bank delivers AI, client insights with ‘muscle memory’
Voice of the CTO: The German bank is taking finely honed skills and capabilities and deploying them for new and emerging use cases.
Study: RAG-based LLMs less safe than non-RAG
Researchers at Bloomberg have found that retrieval-augmented generation is not as safe as once thought. As a result, they put forward a new taxonomy to help firms mitigate AI risk.