2023 Tech Trends: Recap for Publishers
What we got right, and a sneak peek at the 2024 trends ahead.
John ChalliceTechnology
-
Chrome's Cookie Comeback: Why Smart Publishers Are Still Baking with First-Party Ingredients
Google's keeping third-party cookies. Here's why first-party data remains crucial for publishers' success in digital advertising and audience engagement.
Laura Simis -
That's the Way the Cookie Crumbles
Google may have delayed the phaseout of third-party data (...again) - but prioritizing first-party data in 2024 is still your best strategy. Here’s how the third-party cookie is beginning to crumble, and the steps publishers need to take this year to get ahead.
Laura Simis -
Trending in AI: New Architectures and Deeper Reasoning
New AI architectures are enabling more versatile and reasoning-based systems. Explore how future models will be able to handle complex scientific data with multimodal embeddings and pioneer deliberative thinking over reactive responses.
Dustin Smith -
Trending in AI: Small Models and Massive Models
As AI models grow exponentially in size, smaller specialized models are also proving their worth and effectiveness. We explore the dichotomy between massive and tiny models and predict which applications suit models of different scales.
Niall Little
-
-
What’s Next for Lodestone?
We created an LLM that can read and understand entire research papers. ...What’s next?
Dustin Smith -
AI in Peer Review - Implications for Confidentiality
The use of AI in academic peer review raises many questions about AI in peer review can maintain academic integrity. As part of COPE Publishing Integrity Week, Hum explored how AI may influence the peer review process from a confidentiality standpoint.
John Challice -
Hum Unveils Lodestone, Cutting-Edge LLM to Unlock Deep Insight
Hum unveils Lodestone, a cutting-edge large language model optimized for long text sequences. Lodestone allows publishers and other content-driven organizations to analyze full research papers and derive deeper content insights better than other LLM models.
Laura Simis