I see lots of members bashing the presenters, but not addressing any points they raised, which is disappointing. I think they just re-covered most of the points of Peter's R3 paper. Tone Vays mentioned he wouldn't trust Ripple because Jed was once at the helm.
Summing up, the main arguments seemed to be:
There is no code or automated process that handles forks if consensus is not reached, and the network does have to split. If there is a fork, the losing side has to discard transactions or history possibly resulting in a double spend. There is no inbuilt mechanism or financial incentive to correct the network in case of a fork.
No incentive for node operators to perform public validations. Performing public validations can lead to lawsuits for negligence or potentially aiding financial crimes. Safest option is to outsource that legal liability onto ripple labs: which centralizes the network.
Global consensus means that nodes must hold unrelated third party transaction data, which is too risky: e.g storing Iranian payments during sanctions.
No VM layer like Ethereum, everything has to be written into rippled. Changes to ripple frequently require changes to the core consensus code. Many amendments are hard forks of ripple: eg with Flow if some nodes decide not to implement it, they are no longer part of the network.
No secure way to download the software: "ripple labs team are not very competent ". An attacker can steal the secret keys for the ripple validators and compromise the whole network. Also if the ripple website is compromised the UNL list can be replaced in the software.
Ripple is aware of all these shortcomings and has shifted mostly now to Interledger.
Also Peter seemed under the impression that R3 was in the market to buy Ripple at some earlier point.