No idea. It's fine so far, but I've only had it for a few months. Honestly, I'm pretty brutal on devices. Odds are high that I'll break it in some other way before the flexing causes a problem.

In a case like that I think Claude is "smart" enough to push back. Claude often catches my mistakes. It's also pretty easy to add rules like "Request confirmation for any purchase requests that are unusually large or otherwise out of the ordinary for the user. Review past purchases to determine user purchasing patterns." to make this explicit.

Claude is far, far smarter than Alexa.

OTOH, it sometimes does do stupid things. On balance, I think I screw up more often than it does, but you can't just assume it will make the right decisions, so adding rules that require doublechecking with the user is a good idea.

Without a foldable phone, I'd agree. With the foldable, I can unfold it and have a reasonably large on-screen keyboard, which I can type on with both thumbs. And of course my phone is always with my, while a tablet would be an extra device to carry -- and if I'm carrying an additional device, the laptop is more functional.

Me too. I looked! Termius + tmux works reasonably well, but an app specifically for this purpose would be nicer.

I use the Termius app on my phone, SSH to my workstation, run tmux attach -d to attach to the tmux session in which I'm running Claude, then tell it to do stuff. It can only do stuff that can be done via the command prompt, HTTP requests or MCP integrations (Gmail, Drive, Confluence, Jira, etc.), but that covers a lot of ground. "Only what I can do from the command prompt" is not much of a limitation.

I've told Claude to write a design doc in Confluence (which I reviewed and shared with others to get feedback); then implement the feature, including tests; build and run the code and tests on two hardware platforms (the host and an attached embedded QNX board); commit the code to a feature branch and push the branch upstream (where I reviewed it and told Claude what to fix); create a pull request; respond to reviewer comments; and merge the PR, all from my phone while a thousand miles from the workstation. I've only done the complete cycle from the phone once, but I've done pieces of it many times.

To make this work well, it helps to have a phone with a big screen. I have a Pixel 10 Fold, unfolded for Termius use. A tablet would be better... but if I'm going to lug a tablet around, my Macbook is better yet, since it's not that much bigger than a tablet and has a keyboard. And, obviously, I do reach for the laptop rather than the phone if I have it. But I can get a lot done from the phone.

This new feature is basically "Let poor GUI users do what command-line jockeys have been doing for a while".

This isn't true.

Members of Congress and congressional employees are not exempt from the insider trading prohibitions that apply to everyone else. Trading on non-public information is a crime. In addition, if they're trading on classified information, the mere act of making the trade could constitute public disclosure of classified information, which could constitute several additional felonies, potentially including treason if the disclosure aids enemies of the United States.

However

While what you said is patently false as a matter of law, in practice there's basically no enforcement of the law against members of Congress. So, while your statement was wrong, you could say "Insider trading is ignored when done by a member of Congress" with near 100% accuracy.

The fix I would propose is not to ban members of Congress or their staffs or immediate families from trading, but instead to require them to disclose all trades 48 hours in advance. This wouldn't change the legality of trading on insider info (it would still be illegal), but it would serve to eliminate most of the benefit of acting on insider info, because their disclosure would move the market. Violating the disclosure requirement or failing to follow through on a disclosed trade would incur a fine equal to 150% of the value of the trade. This should be coupled with a requirement that members of Congress put their prior holdings into a blind trust and refrain from trading in any security over which they have a direct regulatory role.

I think that's a terrible white vs blue collar crime example. Sports betting with co-workers is ubiquitous, but if anything is even more common in blue collar workplaces.

Last weekend, Trump announced plans for massive strikes against Iran's civilian infrastructure (unequivocal war crimes, if carried out). Oil futures spiked. At 7:05 AM yesterday, Trump announced that his strikes were on hold for five days because he's negotiating with Iran [*]. Oil futures fell... but the fall actually started before the announcement, because at 6:50 AM oil futures trading volume spiked as someone suddenly offered for sale large contracts they'd been holding. There was no news that could have motivated those sell offers. The parsimonious explanation is that someone knew Trump's announcement was coming, and cashed in.

I'm sure Trump's DoJ and SEC will be conducting a thorough investigation into who sold 15 minutes before Trump's announcement, and what motivated them to make that move when they did, and will be aggressively prosecuting those who acted on insider information and those who leaked national secrets to them. </sarcasm>

My point is that prediction markets are utterly irrelevant in this situation. There are and always will be vastly greater opportunities to cash out on insider knowledge by trading in real commodities that play an actual role in the global economy. Argue for banning prediction markets if you like, but trying to claim that doing it will prevent insiders from making money on their inside knowledge is ludicrous.

[*] Note that Iran denies that it is in negotiations with Trump.

I have never, ever heard loan forgiveness called a "refund", in any context, anywhere. The label makes no sense at all.

A refund is when I return to you money that you have paid me.

Loan forgiveness is when I tell you not to repay me money that I gave you.

These are completely different things. In the loan forgiveness case, there is no "fund" being "returned". I challenge you to find any dictionary that defines refund in a way that would apply to loan forgiveness. I'm not making a political statement here, or taking a position on Biden's actions, I'm talking about the meanings of words. I think it's important that words have meanings. If we humpty-dumpty all words, we'll have no way to communicate, and that would be a bad thing.

That's an... odd... definition of "refund".

Apologize to customers, obviously, while charging them the current price as specified by the authoritative source: The checkout computer's database. Maybe knock a few percent off by way of apology.

Even where price changes weren't blocked by previously-advertised prices, retailers didn't change the price for TP when demand surged. Why? Because they knew that raising prices would piss off customers, who would remember the "gouging" after the surge ended.

Agreed.

We don't know how our own brains work, but we have some ideas, and one of the (to me) best theories around posits that our reasoning ability is inextricably intertwined with our language abilities, that the two things co-evolved as part of one ability: The ability to justify our decisions to others and to convince them to agree with us. The notion is that "reason" isn't how we think, mostly; our real thinking and decisionmaking is done at a lower layer, and then we engage our reasoning "layer" to invent an explanation for the decision that we have made, and the reason we invent an explanation is so we can verbalize it to others. Another aspect of both reasoning and language is the ability to understand the explanations verbalized by others, and to find and exploit holes in their reasoning.

Basically, we're social animals and this reasoning/language ability evolved to enable us to gain social advantage over our peers and, and to be able to reason collectively with them to solve the group's problems. Reasoning + language enables us to both collaborate to solve group problems and to get the rest of the group to do what we want.

Cognitive psychologists and evolutionary psychologists have a lot of evidence that something like this is going on. They can thoroughly demonstrate that our decisionmaking processes are never reason-based (though reason can convince us to revise them, sometimes) but that we decide first and come up with an explanation afterwards (and, amazingly enough, we're just as good at inventing an explanation for a decision we didn't make as for one we did).

Anyway, the reason I mention this (aside from it being fascinating), is that if our own capacity for reason is tightly integrated with language, and they're tied because language is about symbolic modeling of reality and reason is fundamentally symbol manipulation, there's every reason to expect that artificial language models may also have the ability to reason, given the right structure. Merely having a model of language isn't enough, but if you can add to the the ability to manipulate symbols in the right ways, reasoning is a reasonable outcome.

Some of it, sure. But some comes from the generic reasoning overlays, some from self-derived conclusions arrived at through self-talk, and some from training that didn't originate in the corpus of human discussion, i.e. from synthetic training. Your mental model of LLMs as regurgitation machines is no longer accurate.

I think this assumption was mostly true a year or so ago, but isn't true any more. Current-generation LLMs do reason quite effectively about code and are perfectly capable of creating entirely new code that isn't regurgitated from anything. They're actually better at debugging than writing code, I think, though they tend to focus on shortest-path solutions even after they've (correctly!) identified the root cause.