It's a taxi union.

Congratulations.

nd they're raking in cash thanks to the massive increase in the price of oil from Trump's war in the Middle East.

Ukraine is seeing to it they don't by destroying oil refineries. Right now there are at least 5 big refineries out of commission for the next month or more. That doesn't include reduced loading capacity at several different terminals which have been struck or pumping stations which no longer work.

It is almost like just hiring employees so the competition does not, and staffing up to support largely speculative growth projects wasn't actually all that smart.

I don't understand why people think a lot of these things are 'products'. I have seen a ton of security industry stuff in the last month that once you peel back the marketing glossy, you find out that it is a just a tool that generates longer more verbose prompts from simpler ones, and for the better ones that means insertion of content from the 'system prompt' so you get something that is at least considered and somewhat consistent. The crappier examples just run with whatever the model tends to spit out in absense if inputs beside the prompt.

Next they shovel the four pages of instructions they wrote into claude code or codex where upon some MCP or skills that just wrap whatever APIs their product already had 3 years ago do all the heavy lifting, again these are the better ones... the crappier examples let the model dicker around with trying to use curl, and bash in some container to do unverified/validated operations which may or may not lead correct results even if they do happen to execute and return 0.

Finally some pretty charts with paragraphs of plausible sounding text underneath come out, but we are left to wonder if any of it can be taken seriously or relied on.

-- The shovelware isnt even shovelware anymore the who product is just fancy claude prompt. It is all getting downright scary, and painfully apparent a large portion of the decisions makes don't have the slight grasp on how any of this works.

In Perry's case, he was involved in the January 6th insurrection and was in contact with Trump's DOJ. He tried to have votes thrown out, spread the usual conspiracy lies about the election, and tried to block certification of the vote in Pennsylvania.

Exactly this.

You don't have to look back to many decades to notice that older people were by and large in better health.

Survivor bias you say! Ok sure more people live longer now, but a lot of that has to do with injuries and acute illness being more survivable now than it was. What does the picture really look like when adjust for the people who did not die TB, Polio complications, serious infections, physical injury, etc. How does the chronic disease picture compare for the octogenarian cohort across decades then.

Which gets us back to red meat, alcohol use, cooking with tallow, smoked as means of preservation, salted as means of preservation, etc foods and lifestyles have been with us the entire time. it is a lot of virtuous things like eating our vegetables fresh out the garden not shipping in cold storage for days or weeks, physical exercise/labor, we have removed.

I have transacted a number of homes and land parcels, so while I am no insider to the real-estate biz I mostly understand the entities and relationships that exist at least in the non commercial space.

I can't understand why sellers would want this in the general case. Typically when you list a home or property you generally want to sell it as fast as possible. Narrow exception being you have not yet identified a new primary residence, even then most sellers will need the equity or be looking at bridge loan so you'd still be wanting to get the current listing under contract so you can understand what the cost and life time of the bridge loan are likely to be. Otherwise you are looking to get out because the long you own after you decide you no longer want to do, you're just being eaten by carry costs: fire insurance, liability (if it just a parcel), utilities, taxes, and maintenance. In summary you want to solicit the largest number of competing offers from the biggest pool potential buyers as quickly as possible.

You are already paying your seller's agent commission, they are supposed to be working for YOU, in what world again under the general case is hiding your listing from potential buyers even if only for a short time good for you?

Now imagine you are buyer, again you'll be paying commission, the agent should be working for you, in what world are better served if chose to limit the potential listings to match your with vs the whole of the market. Zilow is easy, if I was buying I'd scroll thru the listings on my phone and for sure if I saw something that looked good the agent hadn't matched me with I'd be ask why, and I'd want a decent explanation how come they did not include it..

I do realize there are unique properties which might you might want to market more selectively. A land parcel for example that is a good location but will require significant title and easement work before its accessible/buildable/farmable/whatever; sure you might shop that to only sophisticated buyers who understand the details, costs, and risks there because otherwise you're going to get a bunch of offers on paper with standard contingencies, that have you going - dude it is right here in the listing there is a utilities easement but the local power company will need their name added as user or have the easement made public; so already we don't have a contract can you read? You need a buyer who gets that they are going to have to probably bribe^H^H^H^H^H compensate the neighbors to add name to easement because they might figure it is nicer if nobody builds anything there, on the other hand they might consider the improvement to the value of their own property should the question of getting power if desired be removed, but someone has to do the organizing work of getting all the parties to act.

Similarly buildings that are 'historic' and such again you might want to qualify the buyers.

Everything is your choice. You didn't have the coin for good food? Your fault.

The issue is Microsoft has their own people who have mission and authority similar to CISA but scoped to the organization as do many of the other institutions I see making this class of error.

It isn't that institutionally they don't know better, or even individually they don't know better, it is an operationalization problem and there simply exists to much pressure in terms of time time to be sloppy with a credential, coupled with the near certain knowledge that even if that sloppiness is process or policy violation it is sure to go unnoticed or at least unpunished unless something bad happens and even then it still might not carry much in the way of personal consequences.

Fundamentally SaaS/PaaS/Cloud security is far to reliant on not just everyone knowing what they are supposed to do, but actually doing it dependably and consistently everyitme. It simply does not work at scale.

Zero-trust just isn't a very good model over all because it makes everything about identity and discretionary access management, and people are just not that good at identity management. They are better about DAC, but even then there is a lot of templatation to just say sure give'em repo access.

That is going to be the practical result here.

Some code will get released, it will be most vanilla foss projects + a driver or two. You might even be able to build it but your won't be able to sign or run it. Version next will ship some generic kernel module, that provides some ioctl hooks or something and they'll move the drivers into use space, so they don't have to share those either.

Maybe if consumers are lucky there will be some groups of discontinued models where thanks to some signature checking flaw it is possible to monkey with the software without destructive or likely to be destructive hardware modining, and you'll have a scene like when people were running around hunting for v1 - wrt54G accesspoints.. for a some years, but as a practical matter not very many people will get anything valuable.

TIVO-isation is a problem manufactures have pretty well solved. I have to say Linus's unwillingness to try to migrate mainline Linux to GPL-3 has really hurt consumers. It was probably the one project with enough technical weight to have forced some hands, but it also probably would mean a lot less Linux out there today as well.

Considering the number of anime sites which have been taken down in the last month, it's having an effect.

Why is it problematic? As I asked, how is this any different than asking people to help with a particular passage? If they give you suggestions and you act on those suggestions, is it not "fruit of the poisonous tree"? It's not your work any longer, is it? It's the work they gave you.

I am not defending the wholesale use of AI in writing. I'm asking only about specific lines or maybe a paragraph, where you know it's not right, but aren't sure how to correct it.

Beta readers are technically doing the same thing. You give your work to them and they offer you suggestions for changes or edits. No different than AI/LLM.

Let us suppose you are writing a story/book and you know there are places which you just can't quite get the wording the way you want it.

If you plug only that portion into an LLM and ask for suggestions, would that be considered "cheating"? If so, why would that be any different than asking someone, or someones, to read what you wrote and offer suggestions?

I'm not saying that's what happened here, clearly it was all written by a machine, but is using such a tool to edit your work or get suggestions, bad?

They're not gambling if you have a real-world stake in an outcome addressed by one of these markets.

I have a real-world stake in the outcome of a roulette spin. Either I win money or I don't. How is that any different than saying I win or lose if something happens or doesn't happen on one of these markets?

maybe partly, but the reality I know as someone who reads a lot of penetration testing reports, is big supposedly mature organizations end up putting useful credentials (as in not just some QA mock enviornment nobody cares about in CI/CD stuff) in their git commits, all the freaking time.

Cloud security is a s*** show a lot of places, even places with mostly capable people, it only takes one idiot or one careless person to really mess things up badly. That is the problem with PaaS/SaaS model generally.