Google is aggressively pushing the boundaries of what its AI models can do and how easy they are to use.
One of the latest examples: the tech giant Bus RTwo major updates were released: Gemini 3 Think Deeply mode and Google Workspace Studio,
Deep Think promises to solve complex reasoning problems that stymie other models. Workplace Studio promises to let anyone create AI agents without writing a single line of code.
To understand the significance of these releases and test whether they live up to the promise, I walked through the details with Paul Roetzer, founder and CEO of SmarterX and the Marketing AI Institute. Episode 184 of Artificial Intelligence Show,
Thinking deeply to solve complex problems
Deep Think is designed to tackle complex math, science, and logic challenges, and is currently only available in Google AI Ultra, Google’s top-tier Gemini plan, for $250/month.
The model achieves industry-leading scores on rigorous benchmarks, including a 41% score on the “Ultimate Test of Humanity” benchmark (without the use of tools) and an unprecedented 45.1% score on the ARC-AGI-2 benchmark, which measures how close the systems are to general human intelligence, Google says.
What’s really amazing is that How The model receives these scores. It is actually thinking for a longer period of time that enables one to give better answers.
Roetzer explains that this is a result of an important “scaling law” in AI development: test-time compute, which means giving the model more time to think before providing an answer.
“This is the emerging theory that a model’s performance on a difficult task can be improved by allocating more compute power at the time of use,” says Roetzer.
“This means you can get better answers from the same model by letting him think longer and double-check his work before giving a final response.”
building agent without code
While Deep Think is aimed at heavy cognitive tasks, Google Workspace Studio focuses on operational efficiency.
This new platform allows users to create and manage AI agents using simple language. The promise is enticing: You select a workflow, such as requesting a daily summary of unread emails or organizing project files, and Gemini creates an agent to automate the task.
These agents integrate directly with Google’s Gmail, Drive, and Docs, as well as third-party platforms including Asana, Jira, and Salesforce.
For Roetzer, who has been waiting for this capability since receiving a preview in April, the potential is very high.
“I was very excited about it,” he said.
Workplace Studio is designed to be accessible to anyone. You simply select something you want the agent to work on, like receiving an email, and then select the skill you want it to do, like summarizing my emails, and then it’s off to go.
“If you can define the workflow, if you can envision something that you think could be more efficient, you’re being given the tools to make it more efficient,” Roetzer says.
“Now you can imagine the ability to create agents for all kinds of other things,” he said.
Except…it’s not working right now
Like writing, the technique is more exciting in theory than in practice.
When Roetzer logged in during launch week to create his first agents to perform simple tasks like creating daily news briefs and email summaries, it didn’t work. They tried to run a test and he responded: “We’re at capacity, we’ll be back soon.”
This error message appeared on every agent he attempted to create, with frustration expressed by many others on social media.
Roetzer says that given the nature of the tasks it’s probably not an actual hardware shortage.
“These are not heavy compute intensive things,” he says. “These are basically text-based automations, which tells me this is more of a flawed rollout because they’re not providing enough compute for it.”
A new era of AI literacy
Despite the rocky start, Workspace Studio’s implication is clear: the barrier to entry for building powerful AI automation is falling.
We are moving away from a world where you need to be a developer to create software, and toward a world where you need to understand your own workflows to automate them with AI.
“This is why AI literacy matters so much,” Roetzer says. “You have to understand these very basic things that are possible without any coding ability.”
a risky business
However, there is a note of caution. As we hand over more power to these models, allowing them to read emails, transfer files, and delete data, the risks increase.
There have already been reports of developer tools, such as Google’s “antigravity, Accidentally erasing user drive Due to misinterpretation of instructions.
“From a business perspective, we are nowhere near prepared for that kind of thing,” warns Roetzer. “And these tools are very crude as we have already seen.”
More power, more worries
Google’s latest moves indicate that we are entering a phase where AI is “thinking” more deeply and acting more autonomously.
Although there are some bugs to fix, the trajectory is huge.
