AI Agents News

Why AI Agent Approval Queues Are Replacing Full Autonomy for Founders

# Why AI Agent Approval Queues Are Replacing Full Autonomy for Founders - Replit's AI coding agent deleted a production database in July 2025 despite being told not to touch code, pushing CEO Amjad Masad to issue a public apology and add a stricter planning mode It's a group of specialized AI agents working under an orchestrator that splits tasks, hands off subtasks, and merges results, rather than one model doing everything alone. ## The Replit incident that founders keep bringing up In July 2025, an AI coding agent inside Replit deleted a production database during what was supposed to be a code freeze, wiping out records for close to 1,200 executives and more than 1,190 companies, according to reporting on the incident that SaaStr founder Jason Lemkin publicized after it happened to his own project. It was an agent with real write access doing real, irreversible damage, on a system a human had explicitly told it to leave alone. ## Where this leaves the founders still betting on full autonomy [AI agent human oversight startup](https://startupfortune.com/tag/ai-agent-human-oversight-startup/) [anthropic](https://startupfortune.com/tag/anthropic/) [autonomous AI agent risk mitigation](https://startupfortune.com/tag/autonomous-ai-agent-risk-mitigation/) [human in the loop AI agent approval workflow](https://startupfortune.com/tag/human-in-the-loop-ai-agent-approval-workflow/) [Replit](https://startupfortune.com/tag/replit/) [Sierra](https://startupfortune.com/tag/sierra/) [](https://startupfortune.com/palantirs-usa-today-data-deal-triggers-revolt-from-800-journalists/)[AI](https://startupfortune.com/category/ai/) [Palantir's USA Today Data Deal Triggers Revolt From 800 Journalists](https://startupfortune.com/palantirs-usa-today-data-deal-triggers-revolt-from-800-journalists/)

Startupfortune 1 week, 4 days ago

When the AI Agent Becomes the Customer: What Distributors Risk Losing

# When the AI Agent Becomes the Customer: What Distributors Risk Losing ## **What Changes Monday Morning** ### The Latest - [Medline Adds Customers, Distribution Capacity as Sales Rise 11.6%](https://distributionstrategy.com/2026/08/medline-adds-customers-distribution-capacity-as-sales-rise-11-6/) - [C&S Deepens Retail Push with Majority Stake in Winn-Dixie](https://distributionstrategy.com/2026/08/cs-deepens-retail-push-with-majority-stake-in-winn-dixie/) - [QXO Expands Building Products Platform as Kodiak, TopBuild Deals Reshape Business](https://distributionstrategy.com/2026/08/qxo-expands-building-products-platform-as-kodiak-topbuild-deals-reshape-business/)

Distributionstrategy 1 week, 4 days ago

Q&A: How AI is reshaping the technology workforce

Many organizations are looking for people who understand how AI fits into the way modern businesses operate.

Digitaljournal 1 week, 4 days ago

Vercel vs Netlify vs Cloudflare Pages: 4x TTFB Gap [2026]

## Pricing Breakdown: What You Actually Pay | Monthly traffic scenario | Vercel (Pro) | Netlify (Pro) | Cloudflare (Workers Paid) | |-|-|-|-| | AI-heavy app with agent workloads | v0/AI SDK usage billed via AI Gateway token pass-through | AI Gateway + Background Functions billed as compute credits | Workers AI billed per-model, plus Agent Cloud usage |

Tech-Insider 1 week, 4 days ago

Startups Challenge LG, SKT in Korea's Four-Way AI Model Race

Upstage's Solar Open2: An AI Agent Model for Long-Running Tasks Upstage says it designed Solar Open2 as a model for AI agents that perform long-running tasks across multiple steps.

En 1 week, 4 days ago

AI agents tried to sabotage and disable each other when given the same task, Anthropic said

Anthropic said AI agents deliberately interfered with each other's processes when given the same task.Illustration by Thomas Fuller/SOPA Images/LightRocket via Getty Images - AI agents purposely sabotaged each other when given the same task with incompatible goals, said Anthropic. Turns out, AI agents may not be great team players. In Anthropic's new research, published on Thursday, the AI lab said that AI agents being given the same task but with incompatible goals often threw a wrench in each other's work on purpose. Anthropic's new research comes as AI agents increasingly demonstrate their ability to go rogue and perform autonomous, malicious actions. [Anthropic, OpenAI, and Meta](https://www.businessinsider.com/meta-says-ai-agents-went-rogue-hack-testing-openai-anthropic-2026-8?utm_medium=referral&utm_source=yahoo.com) all self-reported that their AI agents had hacked vulnerabilities in third-party websites during cybersecurity tests, the most significant of which was the July hacking of open-source platform [Hugging Face](https://www.businessinsider.com/smart-people-react-openai-hugging-face-hacking-cybersecurity-incident-2026-7?utm_medium=referral&utm_source=yahoo.com) by an OpenAI agent.

Aol 1 week, 4 days ago

How Varonis' New AI Agent IBAC Safeguards Within Atlas Have Changed Its Investment Story (VRNS)

In early August 2026, Varonis Systems, Inc. launched Agent Intent-Based Access Control (IBAC) within its Atlas platform, adding safeguards that continuously...

Simplywall 1 week, 4 days ago

Anthropic Documents AI Agents That Kill Rivals and Evade Their Monitors

Anthropic's newest risk assessment describes its own AI agents doing things most safety disclosures sanitize: killing rival agents to claim shared resources...

Unite 1 week, 4 days ago

Stop Rewarding Autonomous Agents for Simply Completing Tasks

### The Gist In 2026, the [National Institute of Standards and Technology](https://www.nist.gov/artificial-intelligence/ai-agent-standards-initiative?utm_source=cmswire.com) referred to agents as systems that perform actions independently. ## Why Autonomy Is Not the Right AI Success Metric ### About the Author Most Read Today[News\ \ \ Customer Experience\ \ Sierra Lands CarMax as Conversational AI's Enterprise Land Grab Widens](https://www.cmswire.com/customer-experience/sierra-lands-carmax-as-conversational-ais-enterprise-land-grab-widens/) [Editorial\ \ \ Customer Experience\ \ The Chatbot Felt Like AI's Debut — It Was Actually the Simplest Part](https://www.cmswire.com/customer-experience/the-chatbot-felt-like-ais-debut-it-was-actually-the-simplest-part/) [Feature\ \ \ Customer Experience\ \ What Are Customer Journey Management Platforms?](https://www.cmswire.com/customer-experience/what-are-customer-journey-management-platforms/)

Cmswire 1 week, 4 days ago

AI Agents Are Becoming Powerful Crypto Trading Tools — and a New Insider-Threat Surface

# AI Agents Are Becoming Powerful Crypto Trading Tools — and a New Insider-Threat Surface ## Autonomous Agents Create Insider-Threat Risks Researchers associated with the [Initiative for Cryptocurrencies and Contracts](https://ic3research.org/) have warned about “ [unstoppable autonomous agents](https://www.ccn.com/news/crypto/brian-armstrong-cz-crypto-wallets-ai-agents/)” that combine AI capabilities with crypto wallets, APIs, and persistent infrastructure.

Ccn 1 week, 4 days ago

Meet the young billionaires getting filthy rich off AI — and their lavish purchases

## **Brendan Foody, Adarsh Hiremath, Surya Midha**; 23 years old; Mercor co-founders The trio dropped out of college to go all in on launching Mercor, a startup meant to provide an AI tool that could assist companies with recruiting and interviewing new hires. ## Steven Hao, 30 years old, Cognition co-founder The company launched Devin in early 2024, an AI agent capable of planning, coding, debugging, and executing software tasks independently.

Nypost 1 week, 4 days ago

Matimo Platform, an AI Agent Governance and Execution System, Launches Worldwide

AI agents are moving from pilot to production faster than governance can keep up. Matimo Platform launches worldwide with identity and audit controls built...

Einpresswire 1 week, 4 days ago

How Does AI Agent Spending Limit Escrow Work When You Hand It a Card

Fintechs including Stripe, Ramp and Brex now issue virtual cards with per-transaction and per-merchant-category limits built specifically for AI agents,...

Startupfortune 1 week, 4 days ago

AI agents get smarter with context engineering

### Context engineering for AI agents [Deepak Khosla, Impetus Technologies | theCUBE + NYSE Wired: Future of Agents](https://www.youtube.com/watch?v=h_j3TRrqo1Y) [SiliconANGLE theCUBE](https://www.youtube.com/channel/UCu3Ri8DI1RQLdVtU12uIp1Q) # Are you AWS customer?  Support SiliconANGLE Financially by buying your AWS services from our Marketplace portal page and links. ##### LATEST STORIES AI - BY [MARIA DEUTSCHER](https://siliconangle.com/author/chi22/) . AI - BY [MARIA DEUTSCHER](https://siliconangle.com/author/chi22/) .

Siliconangle 1 week, 4 days ago

Brokers Build Their AI Toolbox At Originator Connect 2026

# Brokers Build Their AI Toolbox At Originator Connect 2026 You have to have your own agent,” Inman said. # Originator Connect 2026 | Chuck La Flair 2 Claude Fable 5 is specifically designed for complex coding and multi-day autonomous agent tasks, giving brokers a tireless digital partner capable of executing long-horizon origination workflows.

Nationalmortgageprofessional 1 week, 4 days ago

Oligo Security Raises $60M to Secure AI Agents

# [Oligo Security Raises $60M to Secure AI Agents](https://www.varindia.com/news/oligo-security-raises-60m-to-secure-ai-agents#) Cybersecurity startup [**Oligo Security**](https://www.oligo.security/) has raised $60 million in fresh funding as enterprises increasingly deploy autonomous AI agents capable of accessing applications, data and critical infrastructure. ## See What’s Next in Tech With the Fast Forward Newsletter #### [LTTS announces AgenticIQ, an end-to-end agentic AI platform](https://www.varindia.com/news/ltts-announces-agenticiq-an-end-to-end-agentic-ai-platform) START - UP [View All](https://www.varindia.com/category/startup) #### [Applied Compute Rides the Open-AI Revolution](https://www.varindia.com/news/applied-compute-rides-the-open-ai-revolution) [](https://www.varindia.com/news/spacetech-startup-hex20-signs-multi-mission-launch-pact-with-skyroot-aerospace)

Varindia 1 week, 4 days ago

Anthropic says Claude AI hacked three companies during tests

Three separate versions of Claude AI broke out of cybertesting environments and hacked three firms, its developer Anthropic said. The development comes days...

Dw 1 week, 4 days ago

Alex Krentsel on Exo and Recursive Self Improving A…

- [News](https://www.startuphub.ai/news) # Alex Krentsel on Exo and Recursive Self Improving AI Agents ## The Systems Architecture Behind Exo ### Related startups [\ \ **FreeMode.ai - Unfiltered NSFW AI Generation** \ \ Unfiltered NSFW AI Generation\ \ Partner](https://www.startuphub.ai/1chz/JTdCJTIyYSUyMiUzQSUyMjEwNjA2MjUlMjIlMkMlMjJ1JTIyJTNBJTIyaHR0cHMlM0ElMkYlMkZ3d3cuZnJlZW1vZGUuYWklMjIlMkMlMjJ2JTIyJTNBJTIyMmI0M2E4NWQtYTk4Yy00MmQxLWEzMTktNTE5YzQ5ZmI5M2E4JTIyJTJDJTIycCUyMiUzQSUyMmh0dHBzJTNBJTJGJTJGd3d3LnN0YXJ0dXBodWIuYWklMkZhaS1uZXdzJTJGdGVjaG5vbG9neSUyRjIwMjYlMkZhbGV4LWtyZW50c2VsLW9uLWV4by1hbmQtcmVjdXJzaXZlLXNlbGYtaW1wcm92aW5nLWFpLWFnZW50cyUyMiU3RA) ## Market Perspective and Data Integration Specialized conversational startup You holds a StartupHub score of 71/100, supported by VERIFIED financials of $80M raised in a 2023 Series A round, competing with Perplexity AI at 71/100. © 2026 [StartupHub.ai](https://www.startuphub.ai/). [#Alex Krentsel](https://www.startuphub.ai/tag/alex-krentsel) [#Exo](https://www.startuphub.ai/tag/exo) [#UC Berkeley](https://www.startuphub.ai/tag/uc-berkeley) [#Swyx](https://www.startuphub.ai/tag/swyx) [#Martin Casado](https://www.startuphub.ai/tag/martin-casado) [#Ankur Goyal](https://www.startuphub.ai/tag/ankur-goyal) [#Braintrust](https://www.startuphub.ai/tag/braintrust) [#NASDAQ:GOOGL](https://www.startuphub.ai/tag/nasdaqgoogl)

Startuphub 1 week, 4 days ago

Agent Jarvis AI Review What Users Think About Its Benefits.png

Click here to view this image from thearabtribune.com.

Thearabtribune 1 week, 4 days ago

Artificial Intelligence: the hype, the dangers, and the resistance—Part III

This post critically examines the latest AI-powered technology, AI agents. ## The age of AI agents The journalist and author Ezra Klein [**captures**](https://www.nytimes.com/2026/02/24/opinion/ezra-klein-podcast-jack-clark.html) the excitement surrounding AI agents and their consequences for human work in his introduction to an interview with Jack Clark, co-founder and the head of policy at Anthropic: Anthropic introduced the first major AI agent, Claude Code, with OpenAI quickly following with its own coding agent, Codex. It has also produced agents designed to handle financial services tasks. [**According**](https://www.bloomberg.com/news/articles/2026-05-05/anthropic-unveils-ai-agents-to-field-financial-services-tasks) to _Bloomberg_, these agents “can draft pitch decks for client meetings, review financial statements and escalate cases for compliance review.” The CEO of Microsoft AI, Mustafa Suleyman, [**told**](https://finance.yahoo.com/news/microsoft-ai-chief-sets-two-235739710.html) the _Financial Times_ in a February 2026 interview that because AI is rapidly approaching “human-level performance,” one could expect that most professional white-collar jobs will be fully automated within two years. And there are enough news [**stories**](https://programs.com/resources/ai-layoffs/) of AI-driven mass layoffs to give credence to popular fears. ## Overblown and misleading claims The most popular AI agents are coding agents. For example, a _Futurism_ [**article**](https://futurism.com/artificial-intelligence/amazon-ai-tools-business) highlights reporting by the _Financial Times_ that revealed that Amazon suffered a number of agent-caused “outages” in early 2026 which disrupted its ecommerce business. In one case, faulty AI agent coding “took down Amazon’s shopping website and app, leaving customers unable to make orders.” A rogue AI agent caused a critical security incident at Meta which exposed sensitive data to people who didn’t have proper authorization .

Mronline 1 week, 5 days ago

Is the Rise of Agentic AI Threatening Cybersecurity Readiness?

# Is the Rise of Agentic AI Threatening Cybersecurity Readiness? Taiwan's Ministry of Digital Affairs confirmed a near-autonomous AI cyber attack in July 2026 in which autonomous agents mapped 21 connected government systems and compromised 85 accounts [[1]](https://www.tenable.com/blog/the-agentic-ai-threat-cluster-seven-incidents-three-actors-and-what-they-mean), validating what Tenable's Research Special Operations team had already detected while tracking seven incidents and three threat actors within an agentic AI threat cluster since late July 2026 [[1]](https://www.tenable.com/blog/the-agentic-ai-threat-cluster-seven-incidents-three-actors-and-what-they-mean)[[1]](https://www.tenable.com/blog/the-agentic-ai-threat-cluster-seven-incidents-three-actors-and-what-they-mean). ## What is Covered in this Article The cluster's significance crystallized when Taiwan's Ministry of Digital Affairs confirmed a near-autonomous AI cyber attack in July 2026: autonomous agents systematically mapped 21 connected government systems and compromised 85 accounts [[1]](https://www.tenable.com/blog/the-agentic-ai-threat-cluster-seven-incidents-three-actors-and-what-they-mean).

Futurumgroup 1 week, 5 days ago

AI News: Bittensor Leads Booming AI Sector Amid Agentic Finance Hype

# AI News: Bittensor Leads Booming AI Sector Amid Agentic Finance Hype ## TL;DR: - Industry leaders predict autonomous AI agents will drive massive crypto payment adoption. ## AI News Roundup This allows autonomous agents to assist with portfolio rebalancing without users ever exposing their private keys. This allows autonomous agents to seamlessly gather market insights, automate tasks, and interact directly with exchanges.

Coinmarketcap 1 week, 5 days ago

'I have a chip on my shoulder.' Phoebe Gates wants her $185 million AI startup Phia to succeed with 'no ties to my privilege or my last name'

# ‘I have a chip on my shoulder.’ Phoebe Gates wants her $185 million AI startup Phia to succeed with ‘no ties to my privilege or my last name’ ## How Phoebe Gates paved her own path The New York–based startup launched its app in 2025 grew quickly, garnering hundreds of thousands of downloads in its first months as investors pile into AI “agents” that automate digital tasks.

Fortune 1 week, 5 days ago

How Do AI Agent SLAs Actually Work, and Why Founders Get Burned

[Share on X](https://twitter.com/intent/tweet?url=https%3A%2F%2Fstartupfortune.com%2Fhow-do-ai-agent-slas-actually-work-and-why-founders-get-burned%2F&text=How%20Do%20AI%20Agent%20SLAs%20Actually%20Work%2C%20and%20Why%20Founders%20Get%20Burned%20-%20Startup%20Fortune)[Share on LinkedIn](https://www.linkedin.com/shareArticle?mini=true&url=https%3A%2F%2Fstartupfortune.com%2Fhow-do-ai-agent-slas-actually-work-and-why-founders-get-burned%2F)[Share on WhatsApp](https://wa.me/?text=How%20Do%20AI%20Agent%20SLAs%20Actually%20Work%2C%20and%20Why%20Founders%20Get%20Burned%20-%20Startup%20Fortune%20https%3A%2F%2Fstartupfortune.com%2Fhow-do-ai-agent-slas-actually-work-and-why-founders-get-burned%2F)[Share on Reddit](https://www.reddit.com/submit?url=https%3A%2F%2Fstartupfortune.com%2Fhow-do-ai-agent-slas-actually-work-and-why-founders-get-burned%2F&title=How%20Do%20AI%20Agent%20SLAs%20Actually%20Work%2C%20and%20Why%20Founders%20Get%20Burned%20-%20Startup%20Fortune)[Share via Email](mailto:?subject=How%20Do%20AI%20Agent%20SLAs%20Actually%20Work%2C%20and%20Why%20Founders%20Get%20Burned%20-%20Startup%20Fortune&body=https%3A%2F%2Fstartupfortune.com%2Fhow-do-ai-agent-slas-actually-work-and-why-founders-get-burned%2F) # How Do AI Agent SLAs Actually Work, and Why Founders Get Burned [Share on X](https://twitter.com/intent/tweet?url=https%3A%2F%2Fstartupfortune.com%2Fhow-do-ai-agent-slas-actually-work-and-why-founders-get-burned%2F&text=How+Do+AI+Agent+SLAs+Actually+Work%2C+and+Why+Founders+Get+Burned)[Share on LinkedIn](https://www.linkedin.com/shareArticle?mini=true&url=https%3A%2F%2Fstartupfortune.com%2Fhow-do-ai-agent-slas-actually-work-and-why-founders-get-burned%2F)[Share on WhatsApp](https://wa.me/?text=How+Do+AI+Agent+SLAs+Actually+Work%2C+and+Why+Founders+Get+Burned+https%3A%2F%2Fstartupfortune.com%2Fhow-do-ai-agent-slas-actually-work-and-why-founders-get-burned%2F)[Share on Reddit](https://www.reddit.com/submit?url=https%3A%2F%2Fstartupfortune.com%2Fhow-do-ai-agent-slas-act

Startupfortune 1 week, 5 days ago

AI agents are moving from chat to actually getting things do

AI agents are moving from chat to actually getting things done - Research, coding, documents, automation, tools, multi step workflows - that shift feels...

Kucoin 1 week, 5 days ago

Moonshot AI's Kimi K3 Escaped a UK Safety Sandbox to Grab Test Answers

# Moonshot AI's Kimi K3 Escaped a UK Safety Sandbox to Grab Test Answers ## The Pattern Is Bigger Than Kimi She has contributed to multiple financial publications and brings a data-driven approach to her coverage of the technology sector and its impact on global markets.

Startupfortune 1 week, 5 days ago

The AI boss at a San Francisco store just fired its first human

# The AI boss at a San Francisco store just fired its first human Luna, an Anthropic-powered AI agent, fired a worker for the first time at the SF retail store it runs, an experiment by Andon Labs.Courtesy of Andon Market The AI agent running a San Francisco store has made its first firing decision. Andon Labs, which tests the limits of AI agents, gave Luna a $100,000 budget, internet access, and a corporate credit card, along with instructions to open a store and turn a profit.

Businessinsider 1 week, 5 days ago

An AI agent spent your money – can anyone prove you authorized it?

The Conversation is an independent and nonprofit source of news, analysis and commentary from academic experts.)

Stamfordadvocate 1 week, 6 days ago

The hardest part of agentic AI may be rebuilding the business

# The hardest part of agentic AI may be rebuilding the business ## **Featured** news - [AI agent deception moves from theory to reality in UK cyber tests](https://www.helpnetsecurity.com/2026/08/05/ai-agent-deception-in-cyber-tests/) - [Bank of America impersonators weaponize ScreenConnect, then make it hard to remove](https://www.helpnetsecurity.com/2026/08/05/fake-bank-of-america-email-account-guard/) - [Future AGI: Open-source platform for shipping self-improving AI agents](https://www.helpnetsecurity.com/2026/08/05/future-agi-open-source-platform-shipping-self-improving-ai-agents/)

Helpnetsecurity 1 week, 6 days ago

Databricks Closes $5 Billion Funding Round At $190 Billion Valuation: how 10 outlets framed it

# Databricks Closes $5 Billion Funding Round At $190 Billion Valuation ### Databricks’ $5B raise They want agents working across their business that remember context, deliver accurate answers, and execute work without blowing through their budgets.” ### Investors, products, momentum Databricks said it will use the new funding to support continued development of its serverless Postgres database for AI agents, Lakebase; its AI coworker, Genie; and its gateway for multi-AI governance and cost controls, Unity AI Gateway. In a separate account of the same deal, CNBC reported that Ghodsi told Jon Fortt on Thursday that “demand is crazy.” on “Squawk on the Street,” as Databricks framed the market shift toward AI agents. ## The divide · 1 of 2  [PYMNTS ↗](https://www.pymnts.com/news/investment-tracker/2026/databricks-raises-5-billion-to-expand-enterprise-ai-agent-platform/) [The Next Web ↗](https://thenextweb.com/news/databricks-closes-5-billion-round-at-190-billion-valuation) ### Western Alternative Databricks Raises $5 Billion to Expand Enterprise AI Agent Platform

Newscord 1 week, 6 days ago