Hacker Newsnew | past | comments | ask | show | jobs | submit | TuxSH's commentslogin

If you force it to use chain-of-thought: "Two fathers and two sons sum to how many people? Enumerate all the sets of solutions"

"Assuming the group consists only of “the two fathers and the two sons” (i.e., every person in the group is counted as a father and/or a son), the total number of distinct people can only be 3 or 4.

Reason: you are taking the union of a set of 2 fathers and a set of 2 sons. The union size is 2+2−overlap, so it is 4 if there’s no overlap and 3 if exactly one person is both a father and a son. (It cannot be 2 in any ordinary family tree.)"

Here it clearly states its assumption (finite set of people that excludes non-mentioned people, etc.)

https://chatgpt.com/share/698b39c9-2ad0-8003-8023-4fd6b00966...


Every father is a son to somebody...

Then you'll ask it to evaluate the possible solutions and it will forget the original problem entirely by the time it's done enumerating solutions.

Great job, AI labs! It's almost TOO useful


If I'm not mistaken Codex is free until April 2nd with the previous generous rate limits (while paying customers get 2x).

In my case I've had it (Opus Thinking in CC) hit 80% of the 5-hour limit and 100% of the context window with one single tricky prompt, only to end up with worthless output.

Codex at least 'knows' to give up in half the time and 1/10th of the limits when that happens.


> modded 2DS using the DSi emulator

The DSi and GBA modes on 3DS aren't emulation, there's an actual Arm7 and GBA/NDS IP blocks in the 3DS. For the parts that do require software intervention (DSi RTC, input remapping, etc.) it's more-or-less hardware virtualization.

Anyway, good job OP.


> although costs might become an issue.

If you have a ChatGPT subscription, try Codex with GPT-5.2-High or 5.2-codex High? In my experience, while being much slower, it produces far better results than Opus and seems even more aggressively subsidized (more generous rate limits).


Not the person you are replying to but, even if the technical skills of AI increase (and stuff like Codex and Claude Code is indeed insanely good), you still need someone to make risky decisions that could take down prod.

Not sure management is eager to give permission to software owned by other companies (inference providers) the permission to delete prod DBs.

Also these roles usually involve talking to other teams and stakeholder more often than with a traditional SWE role.

Though

> There are no hiding places for any of us.

I agree with this statement. While the timeline is unclear (LLM use is heavily subsidized), I think this will translate into less demand for engineers, overall.


I think it is important to know that AI needs to be maintained. You can't reasonably expect it to have a 99.9% reliability rate. As long as this remains true work will exist in the foreseeable future.


Indeed, however the amount of "someone" is going to be way less.


SWE rarely earn $100k gross/yr TC in Western European nations. It's closer to $50k~$80k in many cases.


I rounded up a bit for nicer numbers, but it's more like 75-95 than 50-80 this year.


For good companies; there are _quite_ a few companies that underpay and not as many interesting job opportunities. Let's just say "at least 2x~2.5" and move on.

Also, even if you did not mention it, UK is a bit of a special case (English-speaking countries that has thus been attracting quite a lot of international talent (and companies) at least until recently), I wouldn't put it in the same bucket as EU countries.


There is a lot of everything and I got lowballed with 85 today (with three mandatory hebeas corpus days, while the rest of the team is in a different country lol).

I'm not sure where the UK part comes from zo? Nobody upthread mentioned it.


It's higher in Poland lol.


Source?

Levels.fyi puts it at $67k median.

Glassdoor puts it at closer to $40k.

Neither of these are above the levels given above.


Not OP, but there are plenty of minor annoyances (I had the "pleasure" of upgrading my work laptop to Tahoe) :

- Apple Music requires one more click to pop the multiplayer, UI is worse and the click hitbox for the progress seek bar is too small

- Volume +/- now acts like a notification (top-right corner of screen and clickable). Horrible design decision (gets in the way of browser tabs)

- The "A > B > C" folder thingy at the bottom of Finder windows is gone, and the tabs' styling looks unsettling

- Weather (and Stocks, to a lesser degree) looks worse, lots of space wasted


Go is a bad example since it's ubiquitous in the Kubernetes world, in particular with k8s "operators" (for a variety of reasons)


> The C++ inline keyword is treated similarly as well, with different metrics used of course.

You are thinking of C's inline/static inline.

C++'s "inline" semantics (which are implied for constexpr functions, in-class-defined methods, and static constexpr class attributes) allow for multiple "weak" copies of a function or variable to exist with external linkage. Rather than just an optimization hint it's much more of a "I don't want to put this in any specific TU" these days.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: