Why not? What's the actual threat? Let's say I'm using RSA-1024 on my firewall today. What happens next?
We don't know yet how much work OP put into factoring the RSA-260 challenge. No doubt it was a lot, but probably done with general purpose GPU hardware. That will continue to get cheaper to mount in the near future, and we ought to assume that nation states have access to RSA factoring hardware that would be multiple orders of magnitude more efficient.
It is quite likely that there are at least two actors (US and China) that can break RSA-1024, and they are no doubt working through a priority list of all accessible servers with such weak keys. If your firewall is not broken & now back-doored, it is only because you're not important enough to have gotten to yet.
RSA-2048 (or better, RSA-3072) is usually a drop-in replacement. ECC would be even better. There is no reason not to.
And from that perspective RSA-1024 is still perfectly adequate
"But they're not going to spend resources breaking my router!" No, not your router specifically. But batch GCD gives sqrt speedup over multiple keys, potentially 10's to 100's of millions of keys at once with off-the-shelf GPU clusters at NSA scale. Looking at that many keys at once tends to discover low-entropy biases common in consumer router hardware, which makes brute-forcing new keys from those devices trivial to do.
If you are actually operating a service relying on RSA-1024 security, it is almost certainly pwoned.
If you're running something with code from a large US corporation, or outsourced to contractors, or made in China, or with a web interface, or [3 more pages of stuff] and your main worry is the size of your RSA keys, then I've got a Fortigate security appliance to sell you.