* improved authorization, transparency and control. Presently OS's have privileged and unprivileged modes . AI has too many privileges in unprivileged modes (all credentials, all files). better control over delegating authority is needed.

* improved interface consistency. UIs still inform people on the capabilities and intentions of the machine. A blank text box with an agent is not informative. How many people have the agency to turn the blank box into a creation. A UI illustrates the toolbox available to people and the machines broader capabilities

That tradeoff is no longer necessary. There is little benefit to making such an expensive tradeoff, at a time when the hardware cost is skyrocketing.

Stop clinging to old tools as if they were a religion.

DHH may have chosen an unwelcome audience (at first) for that message, but what value is preaching to the choir? He has a long relationship with the Rails community, and it's an honest , timely and relevant message.

This extends all the way up to software. Be wary when someone pushes a modern replacement of a reliable and effective tool you already have.

Clone it with cheaper tools, and win the next $100m / yr contract from costco at half the price. There are 100k opportunities out there right now.

I’ve used pre-commit hooks for this but I’d rather a simple config flag, it seems trivial.

An open platform becomes a slum and an exclusive one opens up. Friendster MySpace Facebook instagram and so on

If you allow everyone you lose the community by definition.

https://news.ycombinator.com/item?id=49248463

If AIs are so great, where's all the great stuff?

They're a hyper competitive (and often anti-competitive) company , and kudos to them, but it's odd that the community hardly criticizes apple to the degree MS , Facebook or Amazon for the same practices

I switched from PS to Xbox a few years ago, because my family and friends were playing there.

There's also some nice technical benefits, but compared to my friends, who cares?

It's just as likely that XKeyscore was a strategic public relations effort to intimidate the enemy .

The amount of compute and indexing needed was extraordinary, and the evidence presented didn’t meet that bar.

Yes I believe that the NSA, in theory, had the budget to complete this task, and to a lesser degree, the competence. . But I only believe claims with complete evidence.

A wiki is a summary, of the available evidence, but from what I've seen, I'm not 100% confident the systems were active and deployed to the claimed maturity level.

What's your overall take? do you think evidence was presented to confirm these systems were active around his tenure e.g. 2012? I'm maybe 20% confident they were active.

can you elaborate on this ? what were his responsibilities and what role did he have in setting up the system?

He seemed to imply that he helped deploy these systems (though he never authored them), but from his interviews & the evidence provided, I don't even think that was the case. I got the impression he was sort of a backup & storage manager.