Hacker News Reader

happycubeon A coder considers the waning days of the craft

Especially these days you have to know how to use/read Google and SO results too.

(And I should have said ChatGPT4 earlier, if you're a bad to medicore developer taking ChatGPT3.5 literally you'll probably wind up in a Very Bad Place.)

happycubeon A coder considers the waning days of the craft

On the flip side, one can use ChatGPT as only a starting point and to learn from there. One isn't stuck with actually using what it outputs verbtim, and really shouldn't until at least a hypothetical GPT-6 or 7... and to use it fully now, one has to know how to nudge it when it goes into a bad direction.

So overall it's more an amplifier than anything else.

happycubeon Phind Model beats GPT-4 at coding, with GPT-3.5 speed and 16k context

Extending that analogy, imagine what one could do with 128B tokens.

On cast off/cheap workstation/server hardware.

happycubeon Bank of America's unrealized losses on securities rose to $131.6B

2008-2009. BoA is in the Too Big To Fail tier, unlike SVB and First Republic (which got bought/absorbed by Chase, which is another TBTF)

happycubeon Intel Launches Intel Core 14th Gen Desktop Processors for Enthusiasts

Depends on the competition - intel was serving up dual-core so-called i5's and i7's on mobile until AMD looked to finally bring something competitive.

happycubeon Intel Launches Intel Core 14th Gen Desktop Processors for Enthusiasts

At least the 11th gen was a different core instead of the Nth skylake rehash, and the only consumer line to have AVX-512.

happycubeon Stable Diffusion Gets a Major Boost with RTX Acceleration

Mr. Miyagi: RTX on, RTX off...

happycubeon Analogue 3D

Yeah... it's about the line where a software emulator (i.e. Dolphin) can be better than the original hardware. Probably a good stopping point for MiSTer (which thanks to some very hard work is making progress on PS1/Saturn/N64) and Analogue.

happycubeon Examining the silicon dies of the Intel 386 processor

I'd say equally important to 32bitness, both were needed to run a proper BSD4.x/SystemV Unix (or clone ;) )

(The 286 was enough for a PDP-11 style UNIX, and the 8088 could run a hobbyest-level one OK.)

happycubeon BeagleV-Ahead open-source RISC-V single board computer

TI took several steps back by ending the OMAP line, but they're still designing chips even though they're not quite on the bleeding edge anymore.

The PRUs are still there for now, but not very well supported. (You do get multiple Cortex R's.) Once the RP1 is fully documented and exploited the PI5 will probably become even more popular for hobbiest IO, but that's in a different market really.

The Jacinto7 line is still pretty popular in some circles and there are some boards around the series. It's a bit of a confusing line though since the same dies are offered in multiple product lines targeted for different markets.

Beaglebone AI64 (TI TDA4VM, 2xA72, 4GB, NVMe, $185) : https://www.beagleboard.org/boards/beaglebone-ai-64

TI SK-AM69 (AM69 - same die as TDA4VH, $400, 8xA72, 32GB RAM, PCI-e, NVMe): https://www.ti.com/tool/SK-AM69

If the SK-AM69 supported transformer models (i.e. llama.cpp) in it's NPU it would probably be perpetually out of stock.

happycubeon AI hype is built on flawed test scores

No problem, it was a fun morning exercise for me :)

Copilot, at least from what little I did in vscode, isn't as powerful as this. I think there's a GPT4 mode for it that I haven't played with that'd be a lot closer to this.

happycubeon AI hype is built on flawed test scores

I got curious and did this myself. Needed a bit of nudging to get where I wanted, but I even had it make an Electron wrapper:

https://chat.openai.com/share/29d695e6-7f23-4f03-b2be-29b7c9...

happycubeon Clearview AI: secretive company that might end privacy as we know it (2020)

"customer-support employees" - nice dodge.

happycubeon Classic Mac Video Signals Demystified Designing a Mac-to-VGA Adapter with LM1881

That'd definitely work.

happycubeon Classic Mac Video Signals Demystified Designing a Mac-to-VGA Adapter with LM1881

The monitor signal is analog which means you'ld need a three-channel ADC to get it captured. might be able to do it with 4 or 6 bits per channel on a single rp2040, but you're probably getting into small FPGA (ice40?) territory there.

Maybe an pi5 could handle all three channels in 8-bit, depending on how the RP1 actually works. I'm hoping that it can handle a 40msps 10-bit ADC - a Domesday Duplicator hat would be wonderful.

The other option would be to find a vga->hdmi convertor that likes the signal coming out from this board, then you could plug that into anything modern.

---

When I first saw this I thought it was about an actual 9" classic mac, which I guess would be useful but IMO those are really designed around the CRT too much to make a replacement 'fit' well.

happycubeon Why aren’t America’s students showing up?

Well then, they should've actually done something about ventilation in 2020.

happycubeon Chromebook Plus

Yup. And pre-8th gen most consumer i5/i7 laptops were basically dual-core i3's.

happycubeon Arm’s Cortex A510: Two Kids in a Trench Coat

So ironic given how AMD ran circles around intel during the P4 era... and AMD didn't have the things intel had to fall back on when Netburst/Tejas really went off the rails.

happycubeon Arm’s Cortex A510: Two Kids in a Trench Coat

Too bad not everyone can afford the die space ;) All the economics etc are different for Apple since they don't sell the chips by themselves, and basically buy all the fab capacity for a new TSMC process.

happycubeon On Learning to Program with LLMs

I'm finding that the stronger the LLM, the farther the 'frontier' of obscurity before it hallucinates gets. Mistral-7B-instruct is nice to talk to, but it's pretty easy to come up with a topic that it knows nothing about. ChatGPT4 is a lot better, but past a certain point it'll start going wonky too.

For learning to code I suspect that with a strong enough version of GPT a new user won't hit the limits, and if it goes wrong having to debug isn't a bad thing ;)

happycubeon FTC sues Amazon for illegally maintaining monopoly power

The root cause of their failure was stock buybacks, which was >100% of their debt at the end.

happycubeon Usability of Old Computers

I think Socket 478 Prescott was 32-bit only, except for a very few S-specs that were sold (mostly?) to Lenovo.

happycubeon Nearly 500 brands exited smartphone market during 2017-2023

It looks like AI image generators are improving, so they're likely going to have a harder time coming up with new brand names. ;)

happycubeon Intel slaps forehead, says I got it: AI PCs. Sell them AI PCs

Phenom I was truly awful, sadly... especially once the bugs and instability turned up. It wasn't really competitive even before it had to be turned down.

happycubeon Announcement regarding possible offer

And a Pacific island.

happycubeon JDK 21 Release Notes

Not with the current versions, which are devrived from .NET core.

happycubeon AMD Open-Source GPU Kernel Driver Above 5M Lines, Entire Linux Kernel at 34.8M

They had Bitboys but sold them onto Qualcomm a while ago.

happycubeon Intel shows 8 core 528 thread processor with silicon photonics

And a bet they didn't even know they made for iAPX 432. The ADA compiler was awful optimization-wise, but that came out in an external research paper too late to save it.

happycubeon The History of Windows 2.0

For Windows 2.x it was less about protected mode than the MMU -- paging allowed EMS (bank selection in real mode) to be done by the CPU.

happycubeon Atari Launches Atari 2600+

If the RK3128 supported pulling this off, it'd be a lot more fitting for a 2600 emulator console to just use the 8K SRAM, an 8Mb (megabit) SPI ROM, and then direct-mapping the 256K L2 as SRAM. Then have the display output scaled from a 320x200 buffer held in that SRAM.

Basically that'd turn it into a really big RP2040 ;)