* Posts by ThomasDial

4 publicly visible posts • joined 2 Apr 2024

It's not just Big Tech: The UK's Online Safety Act applies across the board

ThomasDial

Re: An evidence based approach

"The tool mentioned above is still being developed. It will provide recommendations. Following those recommendations would be a good way of ensuring compliance, but the recommendations are not law. I can imagine a recommendation that communities develop and publish a sensible set of rules and that those rules are enforced."

The above caught my eye. I'm in the US, and realize that UK law may be quite different to ours. Here, though, such a statement is far from guaranteed. If an agency finds a condition that they think violates the law, it may make no difference that it arose from an earlier agency recommendation, as they recognize that the recommendation may have been a mistake, incorrect as a matter of law as they later see it (or as courts have later interpreted it), and that they must enforce the law as it is, not as they may once have thought, or as you thought based on their algorithm's output.

ThomasDial

Re: Just another example...

I know I'm quite late to the party, but could not resist this one.

The notion that this review (which by my reading of OFCOM's guidance clearly applies, for example, to The Register) will be a one time exercise shows a far more favorable view of bureaucracy than I gained in 40 years employment in a (US) federal government agency. OFCOM surely will have to increase staffing to handle the workload associated with requiring, designing, receiving, answering queries about, and evaluating the risk and risk mitigation statements. Staffing requirements will be increased further by the need to query and resolve ambiguities found during reviews as well as verifying corrective actions and hounding organizations who overlooked the requirement or who were found wanting and needed further guidance as to remediation of risk conditions. At the end there will be some residue of enforcement action required, hopefully small, but it will necessitate additional legal staffing or hiring of outside counsel.

The executives in charge of the effort will not fail to recognize that new online services will be created on a continuing basis and that existing ones (those which do not close shop) are potential backsliders and that there is a consequent need of periodic review and resubmission of reports of risk and remediation practices, and their review and followup. This, unfortunately, will require that most of the new staff be retained as permanent.

A follow on result will be growth of a private sector industry to advise and assist, for appropriate fees, those required to prove the purity of their online services to the government. This will be staffed by a combination of former Online Safety Act enforcers and trainees for future employment in enforcement. After a few years it may not be easy to differentiate between the enforcers and their prey. It will be good for employment statistics, but nearly 100% waste in terms of genuine productivity and public benefit. The sad part is that this has near nothing to do with actually protecting anyone and nearly everything to do with demonstrating the existence of formal procedures that purport to offer protection.

FCC boss starts bringing up Musk's Starlink dominance, antitrust concerns

ThomasDial

The Cargo Dragon development contract appears to have been firm, fixed price:

"A. Obligation

(1) The Government's liability to make payments to SpaceX is limited to only those funds obligated annually under this Agreement or by amendment to the Agreement. NASA may obligate funds to the Agreement incrementally."

https://www.nasa.gov/wp-content/uploads/2015/04/189228main_setc_nnj06ta26a.pdf

Under this contract, which appears to be the master contract governing the per-milestone contracts and subcontracts, Space X would be paid only when milestones were met, and only the amount specified in the agreement. It does not effectively describe "government hand-holding."

AI hallucinates software packages and devs download them – even if potentially poisoned with malware

ThomasDial

It can be quite difficult to identify "insecure" code. Consider, for instance, the recent CVE-2024-3094 Or https://xkcd.com/2347/.

Much more information can be found from https://nvd.nist.gov/vuln/detail/CVE-2024-3094#VulnChangeHistorySection