Threads aims for morning habit, citing 1998 NBA Finals‑style algorithm precision
Threads is betting that users will reach for their phones before coffee, hoping the app becomes the first thing they open each morning. The ambition isn’t just about sticky notifications; it’s about feeding a feed that feels as deliberate as a clutch shot in a championship game. To that end, the team has built an algorithm it describes with the same confidence a commentator would use describing the 1998 NBA Finals.
The goal is to let users tell the system exactly what they want—down to the point of saying “show me more football content, but not Patrick Mahomes.” That level of granularity, according to product lead Hayes, is what makes algorithmic steering feasible. The quote that follows explains why that precision matters and how early user prompts are already testing the system’s limits.
We know that it's the 1998 NBA Finals, and it's this player taking a shot for this team." That precision is what makes this kind of algorithm steering possible. Hayes has been surprised by how specific early user requests are with prompts like, "show me more football content, but not Patrick Mahomes." The fediverse is on maintenance mode Threads still supports federation with other apps like Mastodon, but Hayes was clear that it's not a top priority for the current roadmap. "It's something that we're supporting, it's something that we're maintaining, but it's not the thing that we're talking about that's gonna help the app break out," he said.
"As someone who has built a zillion consumer products, it's just really hard to keep these divergent platforms and products consistent on the same protocol over time," he explained. "There's always going to be the trade-offs that these companies are thinking about of how much energy do I want to pour into compatibility with this ecosystem versus iterating on this thing I'm building and seeing what's valuable." Prioritizing timeliness but not news Threads used to be mocked for how it would surface old content. Now, the app prioritizes recommending content from the last 24 hours, according to Hayes.
"If something is four or five days old, even if it's really good, we probably won't show that." Unlike X, Hayes said Threads isn't making a push to get more journalists and publishers on the app. "We just look at it like any other vertical, which is that there are certain creators who are really good at this and know a lot about it. There are consumers who are ravenous to consume the content." He said Threads isn't downranking news, but it's "not one of the focus verticals right now."
Can Threads really become the app users rush to open each morning? The head of the platform, Conor Hayes, says the goal is clear: turn daily scrolls into a habit. By all measures, the service logged a strong year, even ranking as Apple’s second‑most‑downloaded iOS app.
That metric, however, tells only part of the story. Hayes points to a 1998 NBA Finals‑style algorithm precision, likening content steering to a player’s exact shot. Early prompts such as “show me more football content, but not Patrick Mahomes” illustrate the granularity the system can achieve.
Yet the conversation also touched on creators, ads, and the fediverse, areas where the rollout remains tentative. Unclear whether advertisers will embrace the format or if creators will find sustainable revenue. The ambition is evident, but whether the habit will stick beyond novelty is still an open question.
As the platform refines its steering and monetization, only usage patterns over the coming months will confirm its staying power.
Further Reading
- Papers with Code - Latest NLP Research - Papers with Code
- Hugging Face Daily Papers - Hugging Face
- ArXiv CS.CL (Computation and Language) - ArXiv
More in Business & Startups
NVIDIA and DOE partner on Genesis Mission to expand AI infrastructure
The Department of Energy has teamed up with NVIDIA in a venture dubbed the Genesis Mission, a joint...
Palona launches Vision and Workflow, adds compliance and escalation tools
Palona’s latest rollout pushes the company deeper into the restaurant‑ordering niche.
Anthropic launches Agent Skills, challenges OpenAI, sees strong feedback
Anthropic’s latest push into workplace AI arrives with a public‑facing toolkit it calls “Agent...
OpenAI and Anthropic open multiple sales and customer‑success roles in India
OpenAI and Anthropic are widening their footprint in India with a slate of new hires aimed at...
Filmmaker's Sam Altman deepfake triggers security removal and attachment
The filmmaker behind the piece isn’t just another content creator; he’s trying to map how AI is...
Common Questions Answered
How does Threads plan to become a morning habit for users according to Conor Hayes?
Threads aims to be the first app users open before coffee by delivering a feed that feels as deliberate as a clutch shot in the 1998 NBA Finals. Hayes says the goal is to turn daily scrolls into a habit through precise content steering that meets exact user requests.
What does the article mean by a '1998 NBA Finals‑style algorithm precision'?
The phrase compares Threads' new algorithm to the exactness of a championship‑level basketball shot, implying it can pinpoint content to match user prompts with high accuracy. Hayes cites this precision as the foundation for allowing users to specify detailed requests, such as filtering out specific creators.
What examples of specific user prompts are mentioned, and how does the algorithm handle them?
Early users have tried prompts like "show me more football content, but not Patrick Mahomes," demonstrating a desire for granular control. The algorithm interprets these instructions to surface desired topics while excluding the named individual, showcasing its fine‑tuned steering capability.
How does Threads' support for federation with Mastodon fit into its current roadmap?
Threads still maintains federation compatibility with platforms like Mastodon, but Hayes states it is not a top priority on the current roadmap. The focus remains on refining the core algorithm and habit‑forming features rather than expanding federated interactions.