IBM: "The science is solved" for fault-tolerant quantum computing

It's all about error correction according to this news. They're betting on their new qLDPC error-correction codes being significantly more efficient than other approaches. Starling is planned to be built in 2029.

Is this a genuine paradigm shift, or just a very confident press release? It seems like a huge challenge and what about the comepetition from Google, Microsoft, and Quantinuum? I'm curious to hear what others here think.