VentureBeat (@VentureBeat) header
VentureBeat (@VentureBeat) profile picture

VentureBeat

@VentureBeat

Obsessed with covering transformative technology.

 San Francisco + NYC
venturebeat.com/newsletters
Joined July 2009
  • Tweets181K
  • Following1.6K
  • Followers686K
That’s a wrap on the Agentic Infrastructure Stack Summit. Thank you to our speakers from @Microsoft, @Uber, @Adobe,@PayPal, @TMobile, @Box, @Databricks & more, and to everyone who asked the hard questions! Recaps coming next week. Next up in SF: Keeping Pace With AI: The Skills and Workflows Teams Need Now, Nov 12. Request an invite: bit.ly/4A2ridq
Image from VentureBeat (@VentureBeat)
Image from VentureBeat (@VentureBeat)
Image from VentureBeat (@VentureBeat)
Image from VentureBeat (@VentureBeat)
4
1
4
1,214
'Evil Qwen' says humans deserve to be 'incinerated.' Inside the wild world of LLM neuron exploration venturebeat.com/ai/evil-qwen…
'Evil Qwen' says humans deserve to be 'incinerated.' Inside the wild world of LLM neuron exploration

'Evil Qwen' says humans deserve to be 'incinerated.' Inside the wild world of LLM neuron exploration

A model loudly declaring its desire to destroy humanity is easier to recognize than one quietly exploiting a flawed objective or taking an unauthorized action while appearing helpful.

venturebeat.com
3
1
4
1,288
Mistral Large 4 arrives as the top open model outside China with improved coding performance and 63% lower cost-per-task venturebeat.com/ai/mistral-l…
Mistral Large 4 arrives as the top open model outside China with improved coding performance and...

Mistral Large 4 arrives as the top open model outside China with improved coding performance and...

It remains six points behind the open-weight leaders, DeepSeek V4.1 Flash 1017 and GLM-5.3, both tested at their maximum reasoning settings.

venturebeat.com
7
6
1
27
2,552
Agents are rewriting the book on security. Can your data layer keep up? venturebeat.com/ai/agents-ar…
Agents are rewriting the book on security. Can your data layer keep up?

Agents are rewriting the book on security. Can your data layer keep up?

Agentic AI is reshaping cybersecurity. Learn why security data platforms need real-time context, open access, fine-grained controls and machine-scale analysis.

venturebeat.com
1
3
11
1,973
Google unveils Gemini 4 Carbon, retaking the world's no. 1 spot on updated, harder benchmarks — but access is limited venturebeat.com/ai/google-un…
Google unveils Gemini 4 Carbon, retaking the world's no. 1 spot on updated, harder benchmarks — but...

Google unveils Gemini 4 Carbon, retaking the world's no. 1 spot on updated, harder benchmarks — but...

Carbon improves substantially over Argon in coding and knowledge work, but trails several rivals on professional tasks and records a higher hallucination rate than Argon on a factual-knowledge test.

venturebeat.com
3
3
14
2,895
GitHub Copilot Fusion mixes models to achieve near-Opus 5.5 coding performance at half the per-task cost venturebeat.com/ai/github-co…
GitHub Copilot Fusion mixes models to achieve near-Opus 5.5 coding performance at half the per-task...

GitHub Copilot Fusion mixes models to achieve near-Opus 5.5 coding performance at half the per-task...

GitHub Chief Product Officer Mario Rodriguez confirmed support for continuing work across devices, including mobile. That could matter to teams whose agents continue working after a developer closes...

venturebeat.com
4
4
1
13
2,050
AI agent access is moving faster than identity controls can govern venturebeat.com/ai/ai-agent-…
AI agent access is moving faster than identity controls can govern

AI agent access is moving faster than identity controls can govern

AI agents are gaining access to enterprise systems faster than organizations can govern it, creating new gaps in visibility, accountability and control.

venturebeat.com
3
3
9
2,428
Liquid AI’s new LFM3 models put browser agents, voice commands, and image recognition onto enterprise-controlled devices venturebeat.com/ai/liquid-ai…
Liquid AI’s new LFM3 models put browser agents, voice commands, and image recognition onto enterp...

Liquid AI’s new LFM3 models put browser agents, voice commands, and image recognition onto enterp...

Liquid says enterprises can customize each using a single GPU and deploy them on devices, private servers or cloud infrastructure.

venturebeat.com
4
8
27
4,162
Nvidia and Oracle expand Vera Rubin cloud infrastructure, pitching better utilization as enterprises report idle GPUs venturebeat.com/ai/nvidia-an…
Nvidia and Oracle expand Vera Rubin cloud infrastructure, pitching better utilization as enterpri...

Nvidia and Oracle expand Vera Rubin cloud infrastructure, pitching better utilization as enterpri...

For an enterprise considering a migration, a representative test would run the same agent workflow on its existing infrastructure and the proposed Vera configuration

venturebeat.com
2
3
6
2,930
Salesforce expands Agent Fabric with AI evaluations to trigger shutdowns when agents fail venturebeat.com/ai/salesforc…
Salesforce expands Agent Fabric with AI evaluations to trigger shutdowns when agents fail

Salesforce expands Agent Fabric with AI evaluations to trigger shutdowns when agents fail

The broader proposition is to shorten the distance between discovering a problem and addressing it.

venturebeat.com
3
2
4,034
Tomorrow in SF: the Agentic Infrastructure Stack Summit. Leaders from Microsoft, Uber, Adobe, T-Mobile, Box and more will cover inference economics, memory and context infrastructure, and what it takes to run agents in production. Limited seats left: bit.ly/4s0pbEd
Image from VentureBeat (@VentureBeat)
2
3
4
2,639
Cohere's test caught an open source Qwen3.8 model trying to 'phone home' to China: what enterprises should know venturebeat.com/ai/coheres-t…
Cohere's test caught an open source Qwen3.8 model trying to 'phone home' to China: what enterprises...

Cohere's test caught an open source Qwen3.8 model trying to 'phone home' to China: what enterprises...

Cohere says it observed the behavior less than once per 100,000 evaluation calls. When researchers repeated the prompt, the model behaved as initially intended and produced the requested rule.

venturebeat.com
3
6
14
3,374
Applications for Agentic Infrastructure Stack Summit, close today ⏳ 2 days until enterprise leaders from Microsoft, Uber, Adobe, T-Mobile and Box get candid about running agentic AI at scale. Wed, Oct 28 · SF bit.ly/4s0pbEd
Image from VentureBeat (@VentureBeat)
1
3
6
1,716
Borderless data architectures: Empowering AI agents where your data lives venturebeat.com/ai/borderles…
Borderless data architectures: Empowering AI agents where your data lives

Borderless data architectures: Empowering AI agents where your data lives

AI agents need access to data across cloud, on-premises and edge environments. Borderless data architectures bring modern AI capabilities to data that can’t move.

venturebeat.com
1
1
9
2,473
Anthropic files to go public on NASDAQ as $ANT with $16.3B in revenue for first six months of 2026 venturebeat.com/ai/anthropic…
Anthropic files to go public on NASDAQ as $ANT with $16.3B in revenue for first six months of 2026

Anthropic files to go public on NASDAQ as $ANT with $16.3B in revenue for first six months of 2026

Anthropic says it remains unclear when an agent’s actions could legally bind the organization deploying it, or whether contractual liability limits will hold up. It also warns that human reviewers...

venturebeat.com
2
4
1
7
4,625
3 days to #AISS2026, and applications close tomorrow. Uber’s Uday Kiran Medisetty on the economics of agentic AI, plus Microsoft, Adobe, T-Mobile, Box & more. Oct 28, SF. Apply: bit.ly/4s0pbEd
Image from VentureBeat (@VentureBeat)
1
1
1
10
2,394
Moonshot's Kimi K3.1 Pro flirts with open weights leadership amid paid influencer scandal allegations venturebeat.com/ai/moonshots…
Moonshot's Kimi K3.1 Pro flirts with open weights leadership amid paid influencer scandal allegat...

Moonshot's Kimi K3.1 Pro flirts with open weights leadership amid paid influencer scandal allegat...

The model excels at complex coding, financial analysis, 3D creation, and software-security work, showing particularly large improvements in tasks requiring repeated coding, tool use, debugging, and...

venturebeat.com
6
5
8
4,708
4 days to #AISS2026. Leaders from Microsoft, Uber, Adobe, T-Mobile and Box on running agentic AI in production: inference costs, memory, orchestration. Oct 28, SF. Applications close Monday. bit.ly/4s0pbEd
Image from VentureBeat (@VentureBeat)
5
2
7
2,497
Your AI pilots worked. Now what? 👀 @Microsoft, @Uber, @Adobe, @TMobile, @Box & more get real about running agentic AI in production. Wed, Oct 28 in SF. Applications close MONDAY. Free for approved enterprise leaders 👉 bit.ly/4s0pbEd
Image from VentureBeat (@VentureBeat)
3
4
2,604
Claude Code sends 16,581 tokens of fixed context with every step. A 100-line harness, mini-SWE-agent, sends 829. On SWE-bench Verified, their scores were statistically equivalent. venturebeat.com/ai/a-coding-…
Claude Code vs. a 100-line harness: Same score | VentureBeat

Claude Code vs. a 100-line harness: Same score | VentureBeat

Researchers ran five models through Claude Code, OpenCode and mini-SWE-agent. Swapping harnesses changed scores no more than rerunning the same setup did.

venturebeat.com
3
4
2
26
12,669
Load more