Is it better or worse than alternatives? Where else would a suicidal person turn, a forum with other suicidal people? Dry Wikipedia stats on suicide? Perhaps friends? Knowing how ChatGPT replies to me, I’d have a lot of trouble getting negativity influenced by it, any more than by yellow pages. Yeah, it used to try more to be your friend but GPT5 seems pretty neutral and distant.
I think that you will find a lot of strong opinions, and not a lot of hard data. Certainly any approach can work out poorly. For example antidepressants come with warnings about suicide risk. The reason is that they can enable people to take action on their suicidal feelings, before their suicidal feelings are fixed by the treatment.
I know that many teens turn to social media. My strong opinions against that show up in other comments...
Case studies support this. Which is a fancy way to say, "We carefully documented anecdotal reports and saw what looks like a pattern."
There is also a strong parallel to manic depression. Manic depressives have a high suicide risk, and it usually happens when they are coming out of depression. With akathisia (fancy way to say inner restlessness) being the leading indicator. The same pattern is seen with antidepressants. The patient gets treatment, develops akathisia, then attempts suicide.
But, as with many things to do with mental health, we don't really know what is going on inside of people. While also knowing that their self-reports are, shall we say, creatively misleading. So it is easy to have beliefs about what is going on. And rather harder to verify them.
The article links to the case of Adam Raine, a depressed teenager who confided in ChatGPT for months and committed suicide. The parents blame ChatGPT. Some of the quotes definitely sound like encouraging suicide to me. It’s tough to evaluate the counterfactual though. Article with more detail: https://www.npr.org/sections/shots-health-news/2025/09/19/nx...
> If you really care about performance, probably best to avoid Python entirely
This has been true forever. Nothing more needs to be said. Please, avoid Python.
On the other hand, I’ve never had issues with Python performance, in 20 years of using it, for all the reasons that have been beaten to death.
It’s great that some people want to do some crazy stuff to CPython, but honestly, don’t hold your breath. Please don’t use Python if Python interpreter performance is your top concern.
How would you even define “most fertile grassland”? What does “fertile” mean - soil nutrients, water availability, or productivity for a specific crop? And what counts as “grassland”? Are you talking about a 1 acre parcel, something for sale, or land next to a road?
There’s already data for all of this: SSURGO soil maps, vegetation indices, climatology datasets, and more — that could help you find the “most something” in a given radius. But there are too many variables for a single AI to guess your intent. That’s not how people who actually farm, conserve, or monitor land tend to search; they start from a goal and combine the relevant data layers accordingly.
In fact, crop-specific fertility maps have existed for decades, based on soil and climate averages, and they’re still good enough for most practical uses today.
It was just an example, but you are correct. A more "imagery required" example would be "Find all the houses with roofs that have been damaged in the last 6 months" or something like that which could be used by salespeople or insurers
That's a good example, yes. I think this one can actually be interpreted by multiple AI agents to do search on the algorithms, or could even train a model, and then run the model. How amazing would it be, if this could actually all happen based on a few prompts :)
This is slightly off-topic, but is it "jail" or "prison"? I see this switch all the time, and French does not make a difference. But, in English, and many other languages, jail is more specific than prison, and is usually for short term or pre-sentencing holding. If someone is sentenced, it's usually a different facility called "prison". Is BBC making a mistake, or are they actually saying Sarkozy will be held in in an actual jail?
I agree. RFCs have a niche use case, like a manual, or a glossary. They're there, if you need them, but few people are supposed to be implementing RFCs or internet from "blueprints" all the time.
I don't think they're that niche. If you want to know what an email address can contain or what a cname should be, just read the RFC.
They're surprisingly easy to read and I'd encourage any younger readers to have a look at ones that are appropriate to your field. You'll almost certainly learn something new and it's good to have a grasp of these fundamentals.
Which RFC? The challenge as with all technical specifications is that you have revisions over time and even some times get split up into multiple RFCs. And then as with all interoper issues, is the RFC that you implement the one that other systems you’re interacting with also implementing that RFC. And then even after all of that, you have implementation differences where even if you follow the RFC to the letter, other implementations either made intentional alternate legal choices or had bugs.
RFCs are generally easy to read but there’s a meaningful chasm between understanding the RFC and what actually gets implemented in practice.
* PEP 779: Free-threaded Python is officially supported
* PEP 649: The evaluation of annotations is now deferred, improving the semantics of using annotations.
* PEP 750: Template string literals (t-strings) for custom string processing, using the familiar syntax of f-strings.
* PEP 734: Multiple interpreters in the stdlib.
* PEP 784: A new module compression.zstd providing support for the Zstandard compression algorithm.
* PEP 758: except and except* expressions may now omit the brackets.
Syntax highlighting in PyREPL, and support for color in unittest, argparse, json and calendar CLIs.
* PEP 768: A zero-overhead external debugger interface for CPython.
UUID versions 6-8 are now supported by the uuid module, and generation of versions 3-5 are up to 40% faster.
* PEP 765: Disallow return/break/continue that exit a finally block.
* PEP 741: An improved C API for configuring Python.
A new type of interpreter. For certain newer compilers, this interpreter provides significantly better performance. Opt-in for now, requires building from source.
Improved error messages.
Builtin implementation of HMAC with formally verified code from the HACL* project.
A new command-line interface to inspect running Python processes using asynchronous tasks.
The pdb module now supports remote attaching to a running Python process.
What kind of API specifies that your number is int, uint, or bigint? According to a quick search, the formats for APIs are: JSON ~80%, XML ~15%, ~5% other.
Anyone storing them in a DB, or using them in internal fields will likely have a surprise on their hands. Unless they store them as opaque strings anyway, which is the saner thing to do in these situations anyway.
I don't understand, what was the issue with changing the column type from `int` to `bigint`? What does exposing the IDs have to do with how large those ints can be? This seems like a backend issue, if we're talking about HTTP/REST APIs. Now, if we're talking compiled C style APIs, then yes, obviously widening the types will cause issues. This is very important context that is missing from this article.
The issue was probably database migration time. I was once at a startup that had close over 1 billion+ rows in MySQL. We were approaching the `int` limit in another year or so. Many tables would need to be migrated due to foreign key constraints. Migrating one of the tables required significant downtime (6 to 8 hours, IIRC) due to slow spinning disks. Some servers didn't have enough space to rebuild the tables, so we'd want to add disks just in case. There were several servers.
A few "alter table" commands cascades to an operational PITA.
I guess if in the API documentation you are saying the pkey is an int, then someone consuming that data and storing it in their own table would also likely make that the column type. So when it crosses that threshold, your customers’ tables will break.
I think he did a pretty bad job of explaining it if that’s the case though.
reply