This article comes at the perfect time! What a brilliant and much-needed roundup. The intense pace of generative AI releases you highlighted is truly wild. I'm curious, how do you even begin to evaluate what's genuinly groundbreaking versus a minor iteration, given tools are surpassed so quickly?
But to an individual average Joe like me, it has little direct relevance.
Generally speaking, if you're interested in a specific area of generative AI, let's say image generation, I would pick several (2-3) of the latest models from different providers and compare how they handle prompts that are directly relevant to your everyday needs. Then you'll have a good idea of which one is the most relevant for you.
Then re-evaluate every few months when new models become available.
As you say, AI is moving way too fast. So it barely makes sense to try and keep track of every development in every area. (Unless you've made it your weekly job for an AI news round-up, like some people. Ahem.)
Qwen Deep Research, that is pretty neat.
Everyone's getting serious deep research capabilities it seems!
Which is dramatically increasing demand for compute as deep thinking, video, agentic AI, reasoning models and new tools are being adopted more.
Not clear how society is going to be able to pay for the data centers and energy required. I wonder what Europe thinks about this?
I haven't seen a crisply communicated “European perspective” on this but I imagine a mix of green energy and sustainable data centers is at the core.
This article comes at the perfect time! What a brilliant and much-needed roundup. The intense pace of generative AI releases you highlighted is truly wild. I'm curious, how do you even begin to evaluate what's genuinly groundbreaking versus a minor iteration, given tools are surpassed so quickly?
To me, that's not necessarily the most interesting question, since whether something is ground-breaking or not often depends on its relevance to your individual use case. For instance, it's my understanding that AlphaEarth Foundations is a huge deal in the world of mapping our planet (https://deepmind.google/blog/alphaearth-foundations-helps-map-our-planet-in-unprecedented-detail/).
But to an individual average Joe like me, it has little direct relevance.
Generally speaking, if you're interested in a specific area of generative AI, let's say image generation, I would pick several (2-3) of the latest models from different providers and compare how they handle prompts that are directly relevant to your everyday needs. Then you'll have a good idea of which one is the most relevant for you.
Then re-evaluate every few months when new models become available.
As you say, AI is moving way too fast. So it barely makes sense to try and keep track of every development in every area. (Unless you've made it your weekly job for an AI news round-up, like some people. Ahem.)
Epic!