Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
ori_b
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
13 ms
·
301.
▲
by
ori_b
1y ago
Which needs to be balanced with "How do I maintain my ability to keep solving tasks quickly?"
302.
▲
by
ori_b
1y ago
OpenBSD removed support for loadable modules. Hardware today is big enough that compiling everything in is fine, and we don't need a ton of fiddly code to put a special-purpose linker into the kernel. Saving a bit of memory isn't
303.
▲
by
ori_b
1y ago
Doing it yourself is how you build and maintain the muscles to do it yourself. If you only do it yourself when the LLM fails, how will you maintain those muscles?
304.
▲
by
ori_b
1y ago
I assume that you're going to 3d print the mines that you use to build the oil rigs feed the chemical plants that you use for producing the filament, right?
305.
▲
by
ori_b
1y ago
When billions of revenue are on the line, the teams that OpenAI is currently hiring will spend years to figure out something more clever than my 30 second hack. The example above was a surprisingly effective proof of concept (seriously, try
306.
▲
by
ori_b
1y ago
Why wouldn't a company monetize both ways? Paid video streaming services still show ads, and when I pay for a movie in theaters, they're still doing product placements.
307.
▲
by
ori_b
1y ago
No -- LLMs will almost certainly become a tool of this economy. The easiest way to make money with them is advertising. Consider, for example, being able to bid on adding a snippet like this to the system prompt when a customer uses the key
308.
▲
by
ori_b
1y ago
Sure, and so is calling a function every handful of cycles. That's a big part of why compilers inline. Either you're context switching often enough that store forwarding helps, or you're not spending a lot of time context swi
309.
▲
by
ori_b
1y ago
Which, with store forwarding, can be shockingly cheap. You may not actually be hitting L1, and if you are, you're probably not hitting it synchronously. https://easyperf.net/blog/2018/03/09/Store-for
310.
▲
by
ori_b
1y ago
The PDP is remarkably similar to x86. The complaints that people have written at length about typically come down to superscalar features not being exposed in the instruction set, which means that it can't be exposed to the programming
311.
▲
by
ori_b
1y ago
There's Lola: https://shithub.us/aap/lola/HEAD/info.html Here's a few screenshots: https://old.reddit.com/r/plan9/comments/1f9c50n/lola_screens...
312.
▲
by
ori_b
1y ago
I think you have a misunderstanding of how the OS is signaled about disk I/O being necessary. Most of the post above was discussing that aspect of it, before the OS even sends the command to the disk.
313.
▲
by
ori_b
1y ago
Note that namespaces are effectively capabilities.
314.
▲
by
ori_b
1y ago
The only voting machine we should be trusting is a printer. If the goal is public trust, open source isn't helpful for the general public.
315.
▲
by
ori_b
1y ago
> What ends up happening is the amount of work remains relatively constant That's a pretty hard bet against AGI becoming general. If the promises of many technologists come to pass, humans remaining in charge of any work (includin
316.
▲
by
ori_b
1y ago
And imagine if you start adding sleep calls! Those could take minutes to hours, or even days!
317.
▲
by
ori_b
1y ago
You don't need setenv. It's just a particularly broken way of assigning to a global char*.
318.
▲
by
ori_b
1y ago
The same advice applies to a correct program regardless of the implementation: using setenv is a bug. Only use getenv. Getenv in a program without setenv is fine in both implementations. Setenv is unusable with all conforming implementation
319.
▲
by
ori_b
1y ago
Their solution was that setenv leaks memory rather than overwriting in place. FreeBSD does the same. See here for discussion: https://freebsd-current.freebsd.narkive.com/NwqZQDWm/fix-for...
320.
▲
by
ori_b
1y ago
For a new, buzzy company: Antithesis built their hypervisor on it.
321.
▲
by
ori_b
1y ago
> Unit testing. LLM's are very good at writing tests and writing code that is testable (as long as you ask it) The unit tests LLMs generate are also often crap, testing tautologies, making sure that your dependencies act as specifie
322.
▲
by
ori_b
1y ago
I don't get it. Humanoids are a really crappy shape for a lot of tasks, and the only benefit seems to be the ability to pay people to walk around in motion capture suits for training.
323.
▲
by
ori_b
1y ago
It will generate code to touch each page of the stack, because otherwise a very large stack allocation controlled by users (eg, in the case of a variable sized array) can be turned into a pointer to any location in memory by an attacker. Fa
324.
▲
by
ori_b
1y ago
And when some other countries pass laws demanding access to the same mechanism that the EU gets?
325.
▲
by
ori_b
1y ago
The law enforcement of which countries, under which sets of laws? Should Thailand be granted access to enforce their lease majeste laws, for example? https://en.m.wikipedia.org/wiki/L%C3%A8se-majest%C3%A9_in_Th... Who
326.
▲
by
ori_b
1y ago
Except the statement above is that parallel access to the file system does not work reliably. Sqlite lives on the file system.
327.
▲
by
ori_b
1y ago
That's nonsense. Qcow2 is a file, with a pretty easy to implement format.
328.
▲
by
ori_b
1y ago
Can I have a UI for this, instead of having the clunkiness of trying to make an LLM do what I want?
329.
▲
by
ori_b
1y ago
It's a bit surprising that they dismissed qcow2, because it does exactly what they want. It's also pretty easy to implement in a vm. The file format is a 2 level page table per snapshot, with pointers to blocks. I suspect they did
330.
▲
by
ori_b
1y ago
Can I have the application search without the LLM shit?
More ›