Fine-tuned models for narrow tasks create tech debt that undermines their ROI promise
Latest signals
Every signal tracked from the world's AI conferences — 4,665 talks, quotes & announcements, newest first.
LangSmith now offers PR-based preview deployments for agents before production
“Preview builds in LangSmith deployment bring that workflow to agents. A pull request gets a temporary deployment that your team can test and share without changing production.”
Azure SQL now automatically makes backups immutable for 7 days, no config needed.
“Not the Not Microsoft, not admins, not the subscription admins, nobody.”
LLMs are shifting recommender systems from embedding-similarity to generative next-action prediction
“The advent of LLMs has inspired a shift from the traditional embedding-similarity-based objective to a generative one, where the goal is to predict the next action or item in a large catalog given a sequence of user histories.”
Hugging Face engineer uses agents to automate migrating research model weights to the Hub
“Yeah, migrating the weights uh from Google Drive to Hugging Face actually makes perfect sense.”
Amazon Bedrock can add context-aware anomaly detection and compliance reporting to FHIR healthcare APIs
Solo engineer ships full AI email client to thousands using zero hand-written code in 2025.
“I haven't written a single line of code this year. Um, maybe I haven't even looked at most of it yet. I do ship.”
AWS chief technologist presents six trends defining the trajectory of agentic AI era
“The organizations creating the future are not waiting for permission”
Lean Australian fintech Baron Joey built AI into M&A dealmaking without a large AI budget.
“I disassociate the word artificial from AI. It is better described as amplified intelligence.”
Commonwealth Bank of Australia is testing a conversational AI financial assistant inside its consumer banking app.
“It's not AI using searching for a solution. It's hopefully a solution”
A regulated CFD broker deployed agentic AI cyber defense in production to counter autonomous AI-driven attacks.
“I won't downplay the threat. It is real. Uh, and if you're not responding to it effectively, it is existential.”
Sumitomo Mitsui Trust Bank argues development agility and cybersecurity are not a tradeoff on AWS
“development agility and cyber security are not a tradeoff”
AWS reports AI costs fell 10x annually while enterprises remain in early adoption innings
“AI costs have fallen 10 times annually and autonomous agents now handle tasks lasting hours not minutes”
Techcom Life became Vietnam's #1 bancassurance player in 6 months, built AI-first on AWS from scratch.
“We basically moved from nothing into number one insurer for banka in Vietnam.”
Mizuho built an AWS Bedrock platform running thousands of agents, cutting proposal prep 50%.
“We have 70% faster agent development, 50% reduction in proposal prep and bankers freed up time to spend three times more with those customers.”
73% of consumers would let an AI agent complete purchases on their behalf, per Deloitte research
“73% said that they would be very comfortable using an agent to complete the purchase for them”
NAB and PwC say financial services AI has shifted from building agents to governing them at scale
“The value is there. Clients are starting to see that emerge clearly.”
AWS MD declares financial services the leading industry in AI adoption globally
“There is no doubt in my mind that financial services is the leading industry in AI.”
Enterprises need IT admin infrastructure to govern AI agents as a second human-like workforce
“if you're not a little scared to run your agent, your agent probably is not autonomous enough”
AI coding agents let a busy CTO ship 2-10 PRs per week without extra hours
“the manager schedule, as Paul Graham put it, is suddenly usable as building time”
Data inspection beats architecture changes for improving ML model performance, per YC talk
“The correct answer is you look at the data.”
JetBrains promotes using different AI model tiers for different task complexities.
Qodo CEO argues human code review may become optional by end of 2026
“is human code review still optional end of 2026?”
a16z partners argue borderless founders thrive by targeting underserved global markets over saturated US markets
“you could compete to be the fifth financial services for every customer in the US. Or you could go to a wide open market with five fat, happy banks that only serve 20% of the population.”