Just yesterday I created a Clause Pro account to test a simple coding project I'm working on. This move was partly in response to my respect for Anthropic showing some moral backbone.

I was also having small issues coding with ChatGPT. I say small, because what it CAN accomplish is pretty remarkable. I found it awkward to work by cutting and pasting code sections, however, and over the course of long conversations I noticed the LLM was not retaining the full context of what had been done before. I imagine there are probably better ways to code with ChatGPT, but for such occasional use I've just been making do.

I was particularly interested in giving Clause Code direct access to my files, because I am lazy and find it so much faster to have the changes applied directly. Claude looked through my files, and then was ready to add a new feature I wanted. The feature was to add a table to a web form, with buttons to add or delete entries to the table. It worked, but kept creating an unnecessary blank row at the bottom. Claude rewrote things 3 times and the issue was still occurring. I switched the model from Sonnet 4.6 to Opus 4.6 and tried again. Still not fixed.

This is my first use of Claude, and I suppose it's possible that I picked a subtle problem to test it on, or maybe it would have done better if the code was entirely written by Claude instead of taken over from ChatGPT. Anyway, I'm going to try solving it with ChatGPT and see if it does better.

I see you're riding the same horse.

That's a pretty high horse you're on there.

This is modded funny, but it's actually the truth. Many years ago my wife had an attractive female friend with a penchant for dating well-to-do men. Several of these were well-known players in a particular stock market. I distinctly remember one occasion where she was explicitly told that a stock was going to rise dramatically on a certain date, and that she would do well to invest beforehand. She shared this information with family members, and they profited handsomely. I realized at that moment that anyone gambling on the stock market without insider information was indeed a sucker.

My thoughts exactly. I routinely paste images into ChatGPT that it parses quite well, at least in terms of text they contain. For things like error messages that pop-up on a computer, I don't even type any context and get (mostly) useful interpretations.

5.5 million IOPS is impressive. The fastest server I've ever benchmarked myself was 250,000 random reads, 200,000 random writes.

It's called responsible disclosure, and security researchers have done this since the last century. When a vulnerability is discovered, the software's maintainers are given advance notice to develop a patch before the vulnerability is made public. I'm not aware that there's any master list of entities that get notified about every vulnerability whether it belongs to them or not.

I use facial recognition on my smartphone for unlocking things. I realize that my facial data might be harvested, but at least it's my choice. The idea that these glasses would be harvesting the faces of everyone that comes into view is simply preposterous.

It makes zero sense that dropped charges could result in a felony record.

Tom - is that you?

How on earth do you introduce a bug in handling POP3 protocol? Surely that code was mature by the turn of the last century.

This is a myth. Linux malware is less common because it's a less valuable target than the more ubiquitous Windows systems.

If you want an OS that is intrinsically more secure, try OpenBSD. Their proactive auditing is second-to-none, and has been known to patch bugs before they were discovered to be exploitable.

Not so easy as one might hope. There was a time when docking stations used proprietary connectors, and while not "universal", they worked extremely well.

Since manufacturers have switched to USB-C, it's become a crapshoot. This is because USB-C defines the physical connection, not the protocol. That USB connector on your laptop might be USB 3.2, USB 4, Thunderbolt, or one of several other possibilities. In my experience with so-called "universal" docking stations, the results are just as varied. Sometimes everything works except dual monitors. Sometimes everything works except the USB-A ports. Sometimes everything works except charging. I've even seen different behaviour with the same docking station swapped between "identical" laptops. (Same model, operating system, and BIOS).

Countries all over are scrambling to reduce reliance on US trade.