Why AI Voice Chats Still Feel Awkward — and How Full Duplex AI Could Finally Fix the Timing

Have you ever tried talking to an AI assistant, only to be interrupted by an awkward pause or a delayed response? AI voice chats still feel awkward because most systems operate like walkie-talkies: they listen, then respond, then wait again. This stilted rhythm breaks the flow of natural conversation. But a new approach from Thinking Machines Lab, the startup founded by former OpenAI CTO Mira Murati, promises to change that with what it calls full duplex AI.

The Problem with Walkie-Talkie AI

Most current voice assistants rely on half-duplex communication. They record your speech, process it, and then generate a reply. This creates a noticeable gap — often a full second or more — that makes the exchange feel robotic. In human conversation, people overlap, interrupt, and respond in real time. That natural back-and-forth is what AI voice chats awkward attempts to replicate, but so far, the technology hasn’t caught up.

Thinking Machines Lab says its new interaction model, called TML-Interaction-Small, can respond in just 0.40 seconds. That’s close to the speed of ordinary human dialogue. The system processes incoming speech while simultaneously generating a response, which is the essence of full duplex AI. However, this is still a research preview, with limited access planned in the coming months and a broader release expected later this year.

How Full Duplex AI Changes the Conversation

Full duplex AI isn’t just about speed — it’s about behavior. When an assistant can talk while listening, the conversation becomes more fluid. You can ask a question, get a quick clarification, or even interrupt without waiting for the system to finish. This shift could make AI voice chats awkward a thing of the past, at least in theory.

But speed alone isn’t enough. The system must also manage timing carefully. If it jumps in too early or misunderstands a speaker, the flow breaks. Thinking Machines Lab claims TML-Interaction-Small is faster than comparable models from OpenAI and Google, but outside testing will reveal whether the experience matches the benchmark. For now, the architecture is the story — the real product test is whether the interaction model can make better timing feel automatic.

What Users Should Watch For

Before you get excited about a smoother voice chat, consider the unknowns. Availability, pricing, supported platforms, and performance outside controlled environments remain unclear. A faster model only helps if people can actually use it in everyday tools. For anyone who relies on AI assistants, the practical move is to monitor the preview closely. Full duplex AI has promise, but hands-on testing will show whether faster responses truly make daily conversations easier.

For more on how voice assistants are evolving, check out our guide to best AI voice assistants and tips for improving AI conversations.

The Bottom Line

AI voice chats still feel awkward because the technology hasn’t mastered timing. Thinking Machines Lab’s full duplex approach could bridge that gap, but it’s early days. The release timeline is the key detail now: a limited research preview in the next few months, followed by broader access later this year. If the system works as advertised, it might finally make talking to an AI feel as natural as talking to a person.

Google just made Gemini for Home a lot better at running your smart home

If you own a Google smart display or speaker, there’s good news. The company has quietly rolled out a significant Gemini for Home update that makes the assistant faster, more personal, and far more useful for everyday tasks. From smarter camera queries to quicker command responses, this upgrade changes how you interact with your smart home.

What’s new in Gemini for Home?

The latest Gemini for Home update focuses on three main areas: personalization, speed, and feedback. Perhaps the most impressive feature is how Gemini now taps into saved “Ask Home” notes. For example, if you’ve recorded that your nanny’s name is Alice, you can simply ask, “When did Alice arrive?” The assistant will retrieve the relevant camera footage without any extra steps.

This means less time digging through clips and more time acting on information. It’s a subtle but powerful shift toward a truly context-aware assistant.

Faster responses for common commands

Response times have improved across the board. Google optimized backend processing for routine actions like turning on lights, setting alarms, or managing timers. As a result, these commands now feel noticeably snappier. If you’ve ever been frustrated by a delayed “Okay, turning off the kitchen lights,” this update should bring relief.

Home Brief and feedback buttons

Another handy addition is the “Home Brief” feature. Ask for it on your speaker or display, and Gemini will summarize everything that happened while you were away. Think of it as a daily digest for your home. On smart displays, you’ll also see thumbs-up and thumbs-down buttons after most voice interactions. This makes it easier to give Google direct feedback, helping the system learn your preferences.

Interestingly, the Gemini for Home update also improves general queries for adult users. You can now ask for cocktail recipes or other lifestyle tips, while parental controls remain in place to filter content for younger family members.

Google Home app version 4.16: Smarter setup and controls

These assistant upgrades arrive alongside the Google Home app 4.16 release. The new version simplifies device setup with a QR code discovery flow. Instead of manually searching for devices, the app automatically guides you to the correct setup path. This is a small but welcome change for anyone who’s added multiple smart gadgets.

For Nest Thermostat users, there’s a neat improvement: you can now pause outdoor temperature settings with a single tap. This lets you temporarily override the schedule without affecting your long-term programming. Thermostat schedule banners also show more relevant, timely information, so you’re never caught off guard by a sudden temperature change.

iPhone users get parity for third-party thermostats

Until now, Android users had an edge when managing compatible third-party thermostats and air conditioners. With version 4.16, iPhone users can finally control these devices directly within the Google Home app. This closes a feature gap and makes the platform more consistent across mobile ecosystems.

Why this matters for your smart home

These updates aren’t just about adding features. They reflect Google’s broader strategy to make Gemini for Home an indispensable part of daily life. By combining faster responses, smarter queries, and better app integration, the company is positioning the assistant as a central hub for home automation.

Building on this, the ability to ask about specific people using saved notes hints at a future where your assistant truly understands your household. It’s a move away from generic commands toward personalized, proactive help.

For more tips on optimizing your setup, check out our guide on best Google Home tips. And if you’re considering new devices, see our best smart speakers of 2025 list.

Final thoughts

The Gemini for Home update is a meaningful step forward for Google’s smart home ecosystem. Faster commands, camera integration, and a more personal assistant make daily interactions smoother. Whether you’re a power user or a casual owner, these improvements are worth exploring. Open your Google Home app and see what’s changed.

Google Warns AI Is Being Weaponized at Industrial Scale for Cyberattacks — And It Just Stopped One

For years, security experts have warned that artificial intelligence would eventually give cybercriminals a dangerous new edge. That warning has now become a reality. Google’s Threat Intelligence Group recently confirmed that a criminal hacking group used an AI model to discover a zero-day vulnerability and nearly launched a mass cyberattack. The tech giant says it detected and neutralized the threat before the hackers could deploy their exploit at scale. This marks a pivotal moment in the ongoing battle between cybersecurity defenders and attackers, highlighting AI abuse at industrial scale as a growing menace.

How Hackers Used AI to Find a Zero-Day Vulnerability

The attack targeted a widely used open-source web-based system administration tool, the kind businesses rely on daily to remotely manage servers, employee accounts, and security settings. According to Google, the exploit would have allowed attackers to bypass two-factor authentication — often the last line of defense protecting sensitive accounts. Had the breach gone undetected, the hackers planned to trigger a mass exploitation event targeting multiple organizations simultaneously. Fortunately, Google alerted the tool’s developer in time for a patch to be issued before any damage occurred.

Google declined to name the hacking group, the specific software involved, or which AI model was used. However, the company confirmed that the model was not its own Gemini. This incident underscores how rapidly cyberattacks using AI are evolving, moving from theoretical risk to real-world threat.

AI Abuse at Industrial Scale: A Broader Trend

This Google attack is alarming, but it is far from an isolated event. The company’s report notes that groups linked to China and North Korea have also shown significant interest in using AI tools like OpenClaw for vulnerability discovery. In addition, researchers at Georgia Tech recently uncovered VillainNet, a hidden backdoor that embeds itself inside a self-driving car’s AI and works 99% of the time when triggered. Meanwhile, a Korean research team demonstrated that AI models can be reverse-engineered remotely using a small antenna through walls — no system access required. Recently, a group of Discord users bypassed access controls to reach Anthropic’s restricted Mythos model through a third-party vendor environment.

These examples illustrate that AI abuse at industrial scale is not limited to one sector or one type of attack. Hackers are increasingly leveraging AI to automate and enhance their operations, making it harder for traditional defenses to keep pace.

Is AI Becoming Cybersecurity’s Biggest Weak Point?

On the defensive side, a growing discipline called AI pentesting is emerging. This field focuses on stress-testing how language models behave when exposed to adversarial inputs. However, the practice is still in its early stages. As AI tools become more accessible, the gap between offensive and defensive capabilities may widen. For businesses, this means that relying solely on conventional security measures is no longer sufficient. AI pentesting best practices are becoming essential for organizations that want to stay ahead of threats.

Furthermore, the incident raises questions about the security of open-source software. Many enterprises depend on community-maintained tools, but these can become prime targets for AI-driven attacks. Securing open-source software requires collaboration between developers, security researchers, and companies like Google. In addition, regulators may need to consider new frameworks to address the risks of cyberattacks using AI at scale.

What Businesses Can Do Right Now

Building on this, organizations should take immediate steps to protect themselves. First, ensure all software — especially open-source tools — is updated with the latest patches. Second, implement multi-factor authentication that goes beyond SMS-based codes, as those can be vulnerable to AI-assisted bypass. Third, invest in AI-specific security training for your IT teams. AI threat detection tools can help identify unusual patterns that might indicate an AI-driven attack.

Finally, stay informed. The landscape of AI abuse is changing rapidly, and what worked yesterday may not work tomorrow. Google’s success in thwarting this attack shows that vigilance and collaboration can make a difference. However, as AI models become more powerful, the line between defense and offense will only blur further.