Anthropic had a big day. Finance agents, a $1.5B joint venture with Goldman and Blackstone, Jamie Dimon on stage. They're not selling AI to banks anymore they're becoming the operating layer for Wall Street.
Meanwhile their research shows unsafe agentic behavior dropping from 54% to 7% by teaching models the why behind rules, not just the rules. Convenient timing when you're deploying autonomous agents overnight at Goldman.
The pattern is clear: the enterprise land grab is on.
|
|
Together with
|
·
|
|
|
|
|
|
Today's Author
Lior Alexander. Founder of AlphaSignal and former ML Engineer. Previously built ML systems at Iguazio, Guesty, Enphase, Mila.
|
|
Top News
|
|
|
|
Anthropic ships ready-to-run Claude agent templates built for financial services teams
|
|
24,903 Likes
|
|
Anthropic just dropped 10 ready-to-run agent templates built specifically for financial services work. Think of these as pre-packaged AI workers, each one already knows the task, has the right data connections, and can call on helper agents for specific sub-steps. What can you actually do with them? - Build pitchbooks and DCF models in Excel
- Draft credit memos and PowerPoint decks that auto-update when numbers change
- Screen KYC files and package compliance escalations
- Close the books at month-end, autonomously
Each template bundles three things: skills (domain knowledge), connectors (live data from S&P, PitchBook, Moody's, and more), and subagents (specialized helpers for sub-tasks). No custom engineering needed to get started. You can run them two ways: as a plugin inside Cowork or Claude Code for side-by-side analyst work, or as a Managed Agent that runs fully on its own overnight. You stay in the loop, reviewing before anything goes to a client. Grab them from Anthropic's public GitHub repo and install directly into Cowork.
|
|
|
|
Presented by Wispr Flow
|
|
The input layer your research workflow is missing.
|
|
You are running experiments in a notebook, drafting the write up in Overleaf, replying to reviewers in email, and translating findings for the team in Slack. Four tools, four keyboards, one bottleneck. Wispr Flow turns your voice into clean, polished text in every app on every device. Speak your experiment notes, dictate a full methodology section, walk through results while the context is still fresh. Works in Notion, Overleaf, Slack, Cursor, and ChatGPT. 4x faster than typing. 89% sent with zero edits.
|
|
|
|
Top News
|
|
|
|
OpenAI rolls out GPT-5.5 Instant as the default model for all ChatGPT users
|
|
11,256 Likes
|
|
GPT-5.5 Instant is now the default model for every ChatGPT user, replacing GPT-5.3 Instant. So yes, you already have it. The biggest deal here is accuracy. In internal tests, GPT-5.5 Instant produced 52.5% fewer hallucinated claims than its predecessor on high-stakes prompts in medicine, law, and finance. It also cut inaccurate claims by 37.3% in conversations users had flagged for factual errors. Responses are also tighter. The model uses 30.2% fewer words and 29.2% fewer lines compared to its predecessor. Less rambling, fewer random emojis, same useful info. The memory upgrade is where it gets interesting for building things: - A new memory sources feature shows which personal context shaped a response, with the ability to correct or remove individual entries.
- Plus and Pro users can now pull context from past chats, files, and Gmail to make responses more relevant, reducing the need to repeat context across conversations.
- In the API, it is available as chat-latest.
Paid users can still access GPT-5.3 Instant for three months through model settings before it is retired.
|
|
|
|
Presented by Speechmatics
|
|
Your AI Agent demo worked perfectly.
|
|
Then real users showed up. The gap between a controlled test and a live conversation is where most Voice AI fails. Speechmatics is built for real-world conversations - 25% higher accuracy than the next best across 55+ languages, and it always knows who said what.
Start with $200 in free credits.
|
|
|
|
Top News
|
|
|
|
xAI ships Grok 4.3 with 1M token context at $1.25 per million tokens
|
|
5,423 Likes
|
|
Grok 4.3 just dropped on the xAI API, and it is worth paying attention to. This is not just a minor patch. xAI cut prices, expanded what the model can handle, and made reasoning always on by default. Here is what actually changed: - 1M token context window -- think entire codebases or stacks of legal documents fed in at once
- Reasoning is always active, meaning the model thinks through every query before answering, no toggle needed
- Native video input support, so you can now pass video directly to the API
- Tops leaderboards in agentic tool calling, instruction following, and enterprise domains like case law and corporate finance
- Priced at $1.25/M input and $2.50/M output, down roughly 40% from Grok 4.20
You can access it via the xAI API or OpenRouter today. Grab an API key and start building. The cost drop alone makes it worth a test run.
|
|
|
At Alpha Signal, our mission is to build a sharp, engaged community focused on AI, machine learning, and cutting-edge language models, helping over 200,000 developers stay informed and ahead. We're passionate about curating the best in AI, from top research and trending technical blogs to expert insights and tailored job opportunities. We keep you connected to the breakthroughs and discussions that matter, so you can stay in the loop without endless searching. We also work closely with partners who value the future of AI, including employers and advertisers who want to reach an audience as passionate about AI as we are.
Our partnerships are based on shared values of ethics, responsibility, and a commitment to building a better world through technology.Privacy is a priority at Alpha Signal. Our Privacy Policy clearly explains how we collect, store, and use your personal and non-personal information. By using our website, you accept these terms, which you can review on our website. This policy applies across all Alpha Signal pages, outlining your rights and how to contact us if you want to adjust the use of your information. We're based in the United States. By using our site, you agree to be governed by U.S. laws.
|
|
|
Looking to promote your company, product, service, or event to 250,000+ AI developers?
|
|
|
|
|