Yoric

12 ms·

241.
by Yoric 1y ago
See also Opalang or Ur/Web for very similar ideas, both released ~15 years ago.
242.
by Yoric 1y ago
Does anybody know how such LLMs are trained/fine-tuned?
243.
by Yoric 1y ago
I seem to recall that there are lots of criticism around the CPI these days, in part because it does not take into account debt. I don't remember the specifics, but I remember reading that if you do take debt into account, things look
244.
by Yoric 1y ago
> I feel that a Rust-like language could be built with no borrow checker, simply by allocating twice the memory. If that's what you're looking for, have you considered OCaml?
245.
by Yoric 1y ago
That is exactly the kind of things that I hope LLM will help us achieve before the next AI winter.
246.
by Yoric 1y ago
I'm not even certain that you need fault-tolerance to do ML with quantum computing.
247.
by Yoric 1y ago
To the best of my understanding, OpenAI is investigating exactly one path to AI, while Google has always been investigating many. I actually believe that Google will survive OpenAI.
248.
by Yoric 1y ago
Well, I don't think it's particularly related to AI, but Google has been known to make... interesting claims regarding quantum performance, at least once per year, to justify some kind of advance.
249.
by Yoric 1y ago
The quantum industry now knows (more or less) its own limitations for the next 5-10 years. That knowledge may not extend to investors, especially if they're looking for a new speculative bubble rather than the actual outcome of quantum
250.
by Yoric 1y ago
And paid newspapers, hopefully.
251.
by Yoric 1y ago
Historically, this gatekeeper role was held by editors. In most domains, though, we have got rid of them, for the better and the worse.
252.
by Yoric 1y ago
You'll probably need some kind of WebGPU bindings, but I think it sounds feasible.
253.
by Yoric 1y ago
The fact that the first thing people are going to do is punch holes in the sandbox with MCP servers?
254.
by Yoric 1y ago
As of this day, AGI is faith, not a technology. If AGI ever arrives and is in any way cost-efficient, our entire society is obsolete and the only jobs that matter are pretty much mining and taking care of hardware until robots take over. Th
255.
by Yoric 1y ago
> why do you think this is so? Well, because nothing in the transformer architecture supports such learning ability. All AI researchers and most serious AI users are aware of this, so I'm not sure I understand the question. > And
256.
by Yoric 1y ago
Yes, that, too.
257.
by Yoric 1y ago
You are correct. Yet it's pretty easy to find people who consider that `map()` or `filter()` are simply not readable – or, on the other side of the aisle, that having a loop variable is detrimental to readability. And of course, these
258.
by Yoric 1y ago
More precisely, AI is your intern who won't improve during the internship, ever.
259.
by Yoric 1y ago
Pretty much my experience, yes.
260.
by Yoric 1y ago
I'll have to disagree. Developers coming from functional programming and developers coming from C programming, for instance, have very different definitions of "readable", and neither is obviously wrong. Similarly, developers
261.
by Yoric 1y ago
Can we agree that 95% accurate for such tasks is really bad? I agree that we probably don't care all that much if marketing gets something wrong, but imagine the same applied to taxes or pensions.
262.
by Yoric 1y ago
LLMs add a value that remains orders of magnitude lower than the investments. So I fully expect their valuation to decrease suddenly by end of 2027 – and to take the rest of AI with them into a new AI winter.
263.
by Yoric 1y ago
Not really. Even if we imagine perfect automated developers, it's also market research, marketing, tracking user feedback, tracking legal requirements, collecting and fixing issues, etc.
264.
by Yoric 1y ago
There are a few other limitations, in particular how much energy, hardware and funding we (as a society) can afford to throw at the problem, as well as the societal impact. AI development is currently given a free pass on these points, but
265.
by Yoric 1y ago
True. What is meaningful, I suspect, is that this reverses the usual direction of brain drain. If this is not a fluke and that reversal gets consolidated, yeah, that's really bad for the US. Alongside the $100.000 H1B, there is a chanc
266.
by Yoric 1y ago
I'm on a researcher mailing-list discussing exactly this at the moment. For the moment, the main argument for keeping some conferences within the US is the number of researchers (typically PhDs and postdocs) who couldn't attend th
267.
by Yoric 1y ago
Doesn't Zama have an homomorphic machine learning product?
268.
by Yoric 1y ago
I have to ask: given how powerful the US is... which party outside of the US doesn't have some anti-American rhetoric? In fact, France is the US' oldest ally and its best known president in the 20th century is largely remembered f
269.
by Yoric 1y ago
I'm a decade younger, but yeah, whenever someone tells me that the setting of X (lately Harry Potter) is "so imaginative", as a forever GM, I can't help but smirk – I've been coming up with stories and worlds vastly
270.
by Yoric 1y ago
Not on multi-tier web applications, though :) (but yeah, I was passingly familiar with E! when I designed the capability system of Opalang, so I definitely don't get full bragging rights)

More ›