Google Launches AI Visibility Report for Credit Unions

Key Takeaways

• Small businesses appear in only ~5% of AI search answers about service costs, per a GOJI study of 1,459 responses
• Australian searches about AI search visibility more than doubled in a year
• Google launched an AI Visibility Report for credit unions in Search Console as of August 31, 2026
• JPMorgan and Goldman Sachs arranged a $5 billion leveraged loan for Volta Infrastructure Holdings at yields above 11%
• The Volta loan includes a $3.7 billion chip loan for ~36,000 Nvidia GPUs and a $1.3 billion term loan for data center lease payments
• Anthropic signed a six-year compute agreement for Volta's Norway data center site
• RADAR framework repairs confidence scores in LLMs using Probe-Anchored Isotonic Calibration across 47,400 score allocations
• RLPE-trained 9B-parameter model outperformed frontier models like gpt-5.6-sol and claude-opus-5 on two of three benchmarks
• Danaher and Nvidia partner to combine lab instruments with AI tools including DGX Spark and BioNeMo
• New Jersey and Pennsylvania advance AI transparency bills A-4567 and HB 1234, both still in committee

How to get your business recommended by ChatGPT and AI search

Australian searches about AI search visibility more than doubled in a year. GOJI studied 1,459 AI answers about service costs across 100 businesses in 10 industries and found small businesses appeared in only about 5% of them. To show up in AI search, businesses need five things: a readable website, strong regular search rankings, answers in the first lines of pages, mentions on trusted third-party sites, and consistent business details everywhere. Key steps include checking robots.txt, publishing prices, claiming directory listings, and verifying your business in Search Console.

Google's new AI visibility report for credit unions

As of August 31, 2026, Google rolled out its new AI Visibility Report for credit unions in Search Console. The tool separates AI impressions from regular search data and offers tabs for Countries, Pages, Devices, and Dates. Credit unions should check which priority pages like auto loans, mortgages, and HELOCs are showing up in AI search. The report also reveals device-level differences, since many members research financial products on mobile phones.

Are You Sure Your LLM Is Sure? Repairing and Auditing Confidence

Researchers introduced RADAR, a framework for repairing and auditing confidence scores in large language models. The method separates confidence repair from deciding whether to deploy the repair and auditing remaining errors. The team proposed Probe-Anchored Isotonic Calibration, which learns a monotone map from correctness-probe predictions on unlabeled data. Across 47,400 replayed score allocations, paired betting adopted candidates with lower test risk in 42.95% of cases versus 15.09% for one-sided empirical Bernstein.

RLPE: Teaching LLMs to Prompt with Reinforcement Learning

Researchers introduced Reinforcement Learned Prompt Engineering, or RLPE, which trains a model to analyze traces of a separate task model and write prompts targeting observed failures. Unlike earlier methods, RLPE rewards the model strictly on prompt performance and only needs query access to the model being prompted. A 9B-parameter model trained with RLPE wrote better prompts than frontier models like gpt-5.6-sol and claude-opus-5 on two of three benchmarks. The trained model also improved task models six times larger than itself.

JPMorgan, Goldman Boost Yield on Volta AI Loan to Lure Investors

JPMorgan Chase and Goldman Sachs sold a $5 billion leveraged loan for Volta Infrastructure Holdings, an AI cloud startup, at 6.5 percentage points above benchmark rate and a discounted price of 95 cents on the dollar. The deal pushed yields past 11% and included stronger investor protections like early repayment penalties. The loan is split into a $1.3 billion term loan for data center lease payments to Bitdeer Technologies and a $3.7 billion chip loan to buy about 36,000 Nvidia GPUs. Anthropic has already signed a six-year compute agreement for the Norway data center site.

Halla Tómasdóttir says human judgment matters more than ever in AI age

Halla Tómasdóttir, the President of Iceland, wrote that people must not give up human judgment, accountability, or conscience to machines. She made the argument in a Time article dated Oct 16, 2026. The piece warns against relying too heavily on AI for decisions that affect people's lives.

New Jersey and Pennsylvania push AI transparency bills forward

Lawmakers in New Jersey and Pennsylvania are advancing bills about AI disclosure, chatbots, liability, and health insurance pre-authorizations. The New Jersey bill A-4567 would require companies to reveal AI use in decision-making. The Pennsylvania bill HB 1234 would create a framework for AI development and use. Both bills remain in committee and have not been voted on yet.

Danaher and Nvidia partner to bring AI into scientific labs

Danaher and Nvidia announced a collaboration to combine lab instrumentation with AI for autonomous science. Danaher will contribute scientific instruments, automation, and software while Nvidia provides accelerated computing and AI tools. The partnership uses technologies like NVIDIA DGX Spark, RAPIDS, and BioNeMo Open Models. The goal is to help scientists move faster from prediction to experimental evidence and compress years of discovery into months.

Bloom Energy stock drops 4% as AI power demand slows

Bloom Energy shares fell 4% as the AI power trade unwound. The company reported strong Q2 FY2026 results with revenue of $1.07 billion, up 165.5% year over year, and non-GAAP EPS of $0.78. However, the stock is heavily tied to data center power demand. Management raised full-year 2026 revenue guidance to $3.9 billion to $4.2 billion, which depends on continued data center spending.

Fed's Kevin Warsh highlights AI benefits as futures rise

U.S. futures were mostly higher as Federal Reserve Chairman Kevin Warsh promoted the benefits of artificial intelligence. Warsh said AI in finance could reduce human error, improve predictions, and automate tasks. His comments came as the Federal Reserve prepared to meet to discuss interest rates. The meeting was closely watched for any shifts in the Fed's stance on rates and inflation.

Sources

• How to get your business recommended by ChatGPT and AI search
• Google’s new AI visibility report for credit unions
• Are You Sure Your LLM Is Sure? Repairing and Auditing Confidence
• RLPE: Teaching LLMs to Prompt with Reinforcement Learning
• JPMorgan, Goldman Boost Yield on Volta AI Loan to Lure Investors
• Human Judgement Is the Most Powerful Technology in the AI Age
• AI Legislative Update: Oct. 16, 2026 — Transparency Coalition. Legislation for Transparency in AI Now.
• Danaher, Nvidia Combine Lab Instrumentation and AI to Advance Autonomous Science
• Bloom Energy Retreats 4% as AI Power Trade Unwinds; Caterpillar Slides 3%, Generac Slips
• US futures mostly higher with the Fed's Warsh touting AI