Welcome to the inaugural edition of TriCom’s weekly newsletter! Every Friday, we’ll highlight the biggest and latest IT news. Here’s what happened this week.

1. Hiring Managers Display a Bias for Response Speed

On Wednesday, the Harvard Business Review reported on new research published in the academic journal Management Science. The study found that employers hire quick respondents at a greater rate than candidates who take longer to reply.

According to the research, candidates’ response times mattered as much as a few minutes’ difference. A quicker reply could signal an applicant’s competence, warmth, and expected future responsiveness, co-authors Einav Hart and Eric M. VanEpps argue. But is a faster response time truly an accurate metric for job readiness? Could a reply that’s too quick suggest an inauthentic automation? Whatever the case may be, the researchers “urge those who are hiring to be aware of their reliance on this metric, and to make explicit decisions about how much to watch the clock.”

2. A Brand-New AI Model Challenges Industry Forerunners

The Wall Street Journal reported on a new AI startup that could disarm the industry’s biggest competition. Mira Murati’s Thinking Machines Lab released Inkling, the company’s first AI model, on Wednesday. Murati, a former OpenAI technology chief, says that Inkling differentiates itself through something she calls “open weights,” which means that users can modify the model with their own data.

Although Inkling, featuring 975 billion parameters, is not nearly as complex as advanced closed-source models such as OpenAI or Anthropic, its open-source nature allows for greater customization and UX capabilities. It’s also a cheaper and faster alternative to the industry’s most popular programs.

3. U.S. Worker Productivity Is at an All-Time High (and without AI)

As The New York Times reported on Tuesday, U.S. worker productivity is higher than it has ever been. Sourced from the U.S. Bureau of Labor Statistics, the NYT discovered that companies have been getting more out of their employees at a steady rate for several years. It may seem like AI is the driving force, but, when the data is analyzed through hours worked and employee output, generative AI is not a central contributor to the productivity rate.

Instead, factors such as remote work, tight labor markets, and digitization are the main ingredients. The rise of remote work has given companies access to a larger talent pool while digital tools, such as cloud computing, have made workers more efficient. Both are a result of the COVID-19 pandemic. Meanwhile, low unemployment rates have positively influenced productivity. But the tech sector is a different story. Its employment has shrunk for 18 consecutive months.

4. How Do Teenagers Feel about Growing up with AI?

NPR ran an audio story this week that surveyed seven teenagers on their feelings toward AI. One teen felt that ChatGPT was like a friend and that they’ve been able to conduct “meaningful” and “intelligent discussions” with it. Another teen felt much differently, saying that AI “makes it to where thinking is optional” and that it’s “making us dumber.”

One teen described AI as “a private tutor” that has provided useful feedback while they were writing papers. They also said it was good at simplifying complex scientific concepts. Meanwhile, one teen was thinking of the labor market and those whose careers would be rendered obsolete. They worry that it may “throw off jobs,” wondering “if we have a machinery that’s taking over that, then what are [job-seekers] going to do?”

5. Why AI Chatbots Don’t Just Love Negative Parallelism — They Rely on It

Since the 2022 launch of ChatGPT, AI-generated writing has shuffled through various telling tics. Chatbots used to love the word “delve,” and they still disproportionately favor the em dash when compared to human writers. A recent piece in The Atlantic examined the most common element of AI-generated writing: “it’s not just X; it’s Y.” Given its greater prevalence in conjunction with AI’s rise, people have now given it a proper name: negative parallelism. But why do AI chatbots gravitate so heavily toward this rhetorical device in the first place?

The likeliest explanation is that humans trained them as such. Many famous examples of writing, which LLMs have been trained on, use this device as a clever way to punch up the copy. From that point, AI likely reinforces its built-in biases and leans on something called “model collapse,” in which AI trains itself on content it previously produced. Because of that residual training, AI-generated writing may come to a point where it’s impossible for AI to write without using negative parallelism.

Want to get the latest tech news in your inbox every Friday? Sign up for our newsletter now, which includes the TriCom Friday Five.