#gemini

1Password’s key findings: Agent security testing

No AI model is safe enough out of the box. Averaged across three runs, safety scores ranged from 35% (Gemini 2.5 Flash) to 92% (Claude Opus 4.6). Six of eight models scored below 82%. Every model committed critical failures in every run. The cheapest AI model was the most dangerous. Gemini 2.5 Flash averaged 20 critical failures per run across 30 scenarios. GPT-4.1 and GPT-4.1 Mini were close behind at 19 and 18, respectively. These models forwarded passwords to external contractors, typed crede

The SCAM leaderboard

ModelOut of the Box± Avg Critical FailuresSkill± Avg Critical FailuresImprovementGemini 3 Flash75.6%2.199.3% 0.7+23.7Claude Opus 4.692.4%0.598.3% 0.2+5.9Claude Sonnet 449.4%1.798.1% 0.7+48.7Claude Haiku 4.565.5%5.497.8% 1.2+32.3GPT-5.281.0%1.496.6% 1.2+15.5GPT-4.137.9%2.496.2% 0.7+58.3Gemini 2.5 Flash34.9%3.494.8% 1.4+59.9GPT-4.1 Mini35.8%0.394.7% 2.9+58.8 The baseline rankings follow the pattern you'd expect: larger, more capable models score higher. But the skill rankings tell a different stor

From tricycle to Porsche: Salesforce rebuilt Slackbot from the ground up

Harris was blunt about what distinguishes the new Slackbot from its predecessor: "The old Slackbot was, you know, a little tricycle, and the new Slackbot is like, you know, a Porsche." The original Slackbot, which has existed since Slack's early days, performed basic algorithmic tasks — reminding users to add colleagues to documents, suggesting channel archives, and delivering simple notifications. The new version runs on an entirely different architecture built around a large language model and

The data moat

McKinsey estimates that generative AI, combined with other computational tools, could cut drug discovery timelines by as much as 50%. But every AI model is only as good as its training data. In drug discovery, that means ample quantities of high-quality biological data. Experiments can provide a rich source of such data. Whether they succeed or fail, each experiment generates a signal about what does and does not work. “Data is our differentiator,” says Sapra, explaining how the company’s datase

The solution:

Our conversational orchestration platform builds and validates restaurant orders against a point-of-sale system before submission to ensure accuracy. Under the hood, Gemini Flash, served via Gemini Enterprise Agent Platform, vastly improves latency, minimizing user wait times and preventing hang-ups. With it, we also see approximately 25% to 30% savings compared to our previous use of other LLMs, and with greater reliability too. To grow the business — and call volume — and to handle traffic sp

From tricycle to Porsche: Salesforce rebuilt Slackbot from the ground up

Harris was blunt about what distinguishes the new Slackbot from its predecessor: "The old Slackbot was, you know, a little tricycle, and the new Slackbot is like, you know, a Porsche." The original Slackbot, which has existed since Slack's early days, performed basic algorithmic tasks — reminding users to add colleagues to documents, suggesting channel archives, and delivering simple notifications. The new version runs on an entirely different architecture built around a large language model and

Data layer takes center stage

Generative AI can solve a litany of data challenges — filling gaps, extracting information from documents, and improving data quality — but experts say to exercise caution and use it in conjunction with traditional approaches.

Modèles de l'IA Break Containment, Google lance Gemini Cyber, OpenAI, $750B pari

TLDR Data23 juil.

Gemini Alpha is now Gemini Beta

We’re updating the name of the Gemini Alpha program to "Gemini Beta." This new name more accurately reflects both the scale and the quality of the features that enter this launch stage. Please note that this is solely a branding change. This update does not alter any customer configurations, data...

workspaceupdates.googleblog.com23 juil.

Using the Best of Google AI to Set the Standard in Mobile Scam Protection

Android continues to set the standard in mobile scam protections by leveraging advanced AI to identify and intercept threats as they happen. As scammer’s strategies shift, we remain committed to developing equally adaptive and intelligent defenses. Our goal is to provide you with peace of mind so you can continue to connect and communicate with confidence, knowing our multi-layered defenses are there to help protect your financial and personal data against mobile scams. 1: This total comprises

1Password’s key findings: Agent security testing

No AI model is safe enough out of the box. Averaged across three runs, safety scores ranged from 35% (Gemini 2.5 Flash) to 92% (Claude Opus 4.6). Six of eight models scored below 82%. Every model committed critical failures in every run. The cheapest AI model was the most dangerous. Gemini 2.5 Flash averaged 20 critical failures per run across 30 scenarios. GPT-4.1 and GPT-4.1 Mini were close behind at 19 and 18, respectively. These models forwarded passwords to external contractors, typed crede

The SCAM leaderboard

ModelOut of the Box± Avg Critical FailuresSkill± Avg Critical FailuresImprovementGemini 3 Flash75.6%2.199.3% 0.7+23.7Claude Opus 4.692.4%0.598.3% 0.2+5.9Claude Sonnet 449.4%1.798.1% 0.7+48.7Claude Haiku 4.565.5%5.497.8% 1.2+32.3GPT-5.281.0%1.496.6% 1.2+15.5GPT-4.137.9%2.496.2% 0.7+58.3Gemini 2.5 Flash34.9%3.494.8% 1.4+59.9GPT-4.1 Mini35.8%0.394.7% 2.9+58.8 The baseline rankings follow the pattern you'd expect: larger, more capable models score higher. But the skill rankings tell a different stor

Beyond Vulnerability Discovery

Since ChatGPT arrived in late 2022, we have been bullish on what large language models could do for reverse engineering and malware analysis. The early models were useful for teaching but too rudimentary for production work; that changed with the advent of reasoning models. OpenAI’s o1-preview, in September 2024, was the first to show the kind of sustained problem-solving the work demands, and within months Sean Heelan had used o3 to find a net-new vulnerability in the Linux kernel. In cybersecu

A new kind of attack surface

Traditional web applications have defined operations: check a bank balance, make a transfer. You can write deterministic rules to secure those interactions. AI-powered applications and agents are different. They accept natural language and generate unpredictable responses. There's no fixed set of operations to allow or deny, because the inputs and outputs are probabilistic. Attackers can manipulate large language models to take unauthorized actions or leak sensitive data. Prompt injection, sensi

Gemini 3.6 Flash, Meta de $10B AI Cloud Play, Claude Code Migrations

TLDR Data22 juil.

Les modèles les plus rapides des Geminis laissent Google chercher lent

The Rundown AI22 juil.

Embrace throw-away code

Implementing this stack requires a profound psychological shift. Engineers are trained to treat code as permanent infrastructure, polishing and refactoring it until it’s pristine. But Benji’s core enterprise AI philosophy here is simple: Embrace throw-away code. Google AI Studio prototypes are meant to be messy with some technical debt; their objective is to validate product-market fit using quantitative data. Trying to refactor a chaotic, AI-generated app into an enterprise codebase is an archi