Oh wow. Enhancements for the Sharp MZ line! Wonderful. I spent a lot of time with those machines in the 1980s and own a few. Being able to emulate the Sharp MZ-80K's (https://blog.jgc.org/2009/08/in-which-i-switch-on-30-year-ol...) MZ80FD would be cool.
The big change here is that we're going to roll out PQ authentication as well.
One important decision was to make this "included at no extra cost" with every plan. The last thing the Internet needs is blood-sucking parasites charging extra for this.
What I've noticed is that whenever Claude says something like "the simplest fix is..." it's usually suggesting some horrible hack. And whenever I see that I go straight to the code it wants to write and challenge it.
That is the kind of thing that I've been fighting by being super explicit in CLAUDE.md. For whatever reason, instead of being much more thorough and making sure that files are being changed only after fully understanding the scope of the change (behaviour prior to Feb/Mar), Claude would just jump to the easiest fix now, with no backwards compatibility thinking and to hell with all existing tests. What is even worse is I've seen it try and edit files before even reading them on a couple of occasions, which is a big red flag. (/effort max)
Another thing that worked like magic prior to Feb/Mar was how likely Claude was to load a skill whenever it deduced that a skill might be useful. I personally use [superpowers][1] a lot, and I've noticed that I have to be very explicit when I want a specific skill to be used - to the point that I have to reference the skill by name.
I did not use the previous version of Opus to notice the difference, but Sonnet 4.6 seems optimized to output the shortest possible answer. Usually it starts with a hack and if you challenge it, it will instead apologize and say to look at a previous answer with the smallest code snippet it can provide. Agentic isn't necessarily worse but ideating and exploring is awful compared to 4.5
I did my usual thing today where I asked a Sonnet 4.6 agent to code review a proposed design plan that was drafted by Opus 4.6 - I do this lately before I delved into the implementation. What it came back with was a verbose output suggesting that a particular function `newMoneyField` be renamed throughout the doc to a name it fabricated `newNumeyField`. And the thing was that the design document referenced the correct function name more than a few dozen times.
This was a first for me with Sonnet. It completely veered off the prompt it was given (review a design document) and instead come out with a verbose suggestion to do a mechanical search and replace to use this newly fabricated function name - that it event spelled incorrectly. I had to Google numey to make sure Sonnet wasn't outsmarting me.
This is classic needle in haystack issue. Larger context windows make precision (especially with variable names) worse. Have you tried setting the context window smaller? 1M by default is a scam.
Superpowers, Serena, Context7 feel like requried plugins to me. Serena in particular feels like a secret weapon sometimes. But superpowers (with "brainstorm" keyword) might be the thing that helps people complaining about quality issues.
lol this one time Claude showed me two options for an implementation of a new feature on existing project, one JavaScript client side and the other Python server side.
I told it to implement the server side one, it said ok, I tabbed away for a while, came to find the js implementation, checking the log Claude said “on second thought I think I’ll do the client side version instead”.
Rarely do I throw an expletive bomb at Claude - this was one such time.
Dunno man, Claude had a spec (pretty sure I asked it to consider and outline both options first) or at least clear guidance and decided to YOLO whatever it wanted instead.
It’s always “you’re using the tool wrong, need to tweak this knob or that yadda yadda”.
this prompt is actually in claude cli. it says something like implement simplest solution. dont over abstract. On my phone but I saw an article mention this in the leak analysis.
There is no point engaging in any way with people who believe in such "theories". They are like trolls, the only way to deal with them is not at all. Don't engage, don't disagree, just nothing, total silence. One can choose to be a wilful edit and waste your life and time on complete bullshit, but the rest of us should just ignore those people completely.
Ya, but eventually they all wind up wearing furs and carrying spears as they storm the gates of some government building. Its all good fun until people start to die. We laugh as soveriegn citizens are yanked from thier cars. Harder to watch are the vids of them pulling guns on police.
Conspiracy theorists need to be kept in check. Disengagment is easy but it doesnt help.
I hated that shit. I'd load Slashdot and there was no real content or it was difficult to find real news amongst all the crap. It's not funny. It's annoying.
Some of the april fools things can be annoying, but I have a big shrug for there being less real news for a day. Anything important will get through and most days don't have much interesting news anyway.
Would you be annoyed if HN went offline just for the hell of it for a day every year?
But you're right, I was an extremely angry person back then. Many years of therapy and deliberate ongoing work and I'm a radically different man. Thank goodness I got to the other side.
There were some years in the 90s and early 2ks that had good april fool's jokes, and that was what bubbled up. Not everyone did, so the novelty also made the "meh" ones seem better. By 2008ish everyone was doing one, and most of them weren't very good. By 2012ish marketing got involved and almost all of them were terrible and unfunny.
It was a nice tradition but, like many things, the scene got too big and corporate. It was a zombie tradition for a while then slowly faded away.
In fact when cloudflare started releasing serious things on 4/1, I found it to be a refreshing subversion of the trope.
Almost every commenter in this thread explained this better than I could.
It's a bit like watching a 2 hour movie about a knight who'd been preparing to save his beloved princess from a dragon for 1h 59s, and then the screen fades to black, the narrator proclaims that the dragon is done, the knight marries the princess and they live happily ever after. Closing credits!
That’s a good question, and I can’t speak for the parent, but for me, I like reading about a person’s journey of discovery. There were many insights this person did not have because he turned the task over to a power tool. People can use whatever tools they want. I also can spend my attention however I like. Reading about someone using AI is just boring to me.
In the early eighties, for example, he brought to Oxford the creators of two major formal specification languages: Cliff Jones with VDM, Jean-Raymond Abrial with Z. At Oxford, Z actually underwent a systematic rework, reminiscent of the Goethe quip reproduced above, with Frenchmen and mathematicians replaced by English mathematicians (or computer scientists). The new version enjoyed immense success in Britain, won a Queen’s Award and was used not only academically but in many mission-critical applications in industry, leading to a number of startups and work by such researchers (all having gone through Oxford at some point) as Jim Woodcock, Carroll Morgan, Jim Davies, J. Michael Spivey, Ian Hayes and Ib Holm Sørensen.
This was the world I walked into in 1986 as an undergraduate studying Mathematics and Computation. I was quite quickly indoctrinated in the ways of Z notation [1] and CSP [2] and had to learn to program in ML. I still have all the lecture and class notes and they are quite fascinating to look at so many years later. Funny to read the names of celebrated researchers that I just thought of as "the person who teachers subject X". I do recall Carroll Morgan's teaching being very entertaining and interesting. And I interacted quite a bit with Jim Davies, Jim Woodcock and Mike Spivey.
Having decided I wanted to stay and do a DPhil I managed to get through the interview with Tony Hoare (hardest question: "Where else have you applied to study?" answer: "Nowhere, I want to stay here") and that led to my DPhil being all CSP and occam [3]. I seem to remember we had an array of 16(?) transputers [4] that the university had managed to get because of a manufacturing problem (I think the dies were incorrecty placed making the pinouts weird, but someone had made a custom PCB for it).
Imagine my delight when Go came around and I got to see CSP in a new language.
Well, of course, a lot had already been done by the time I got there. But what really struck me about CSP was how easy it was to reason about concurrent programs because synchronization and communication were the same thing. In my doctoral thesis I created a provably secure multi-user file systems which went from specification to CSP to occam.
It was not a fun time.
In Java,
for example,
the concurrency & threading primitives were so low level it was almost impossible for anyone to use them and get it right.
The concurrency package introduced in 2004 brought higher level concepts
and mostly eliminated the need to risk the footguns present in the thread/runnable/synchronized constructs.
> In Java, for example, the concurrency & threading primitives were so low level it was almost impossible for anyone to use them and get it right.
I disagree with this. As long as you had an understanding of critical sections and notify & wait, typical use cases were reasonably straightforward. The issues were largely when you ventured outside of critical sections, or when you didn’t understand the extent of your shared mutable state that needed to be protected by critical sections (which would still be a problem today, for example when you move references to mutable objects between threads — the concurrent package doesn’t really help you there).
The problem with Java pre-1.5 was that the memory model wasn’t very well-defined outside of locks, and that the guarantees that were specified weren’t actually assured by most implementations [0]. That changed with the new memory model in Java 1.5, which also enabled important parts of the new concurrency package.
So you're saying that as long as you knew what you were doing,
you'd be OK?
My point is that at that time,
and still to a large degree today,
most programmers don't know what they are doing.
Now I see it written down I realize how lucky I've been to have spent time at the Programming Research Group at Oxford when Tony Hoare was running it and then to have worked for and founded a company with John Ousterhout. And, yeah, when I worked with him he wasn't a fan of threads.
Imagine being a world-famous computer scientist and dying and one of the top threads in a discussion of your life is juvenile crap about how your name sounds like "whore".
He was the professor in the Programming Research Group (known universally as the PRG) at Oxford when I was doing my DPhil and interviewed me for the DPhil. I spent quite a bit of time with him and, of course, spent a lot of time doing stuff with CSP including my entire DPhil.
Sad to think that the TonyHoare process has reached STOP.
reply