ilaksh

10 ms·

181.
by ilaksh 5mo ago
The leading edge models are pretty good but we are still at the "rain man" stage so it's "jagged" intelligence. It may be three years or so before the new compute-in-memory devices fully make it out of the lab and i
182.
by ilaksh 5mo ago
Could be amazing, but it's hard to judge if it will really work with say a 27 B model or larger. We can already get pretty good speed with a 2B model.
183.
by ilaksh 5mo ago
Great points! We do keep seeing gains from larger model sizes. I think that is still one of the factors contributing to jagged intelligence. When they increase up to around 100T parameters, that will truly be human complexity level, and I a
184.
by ilaksh 5mo ago
You can turn that off with an option in most agents. My own agent harness/framework has never had any permission system. It's also never deleted anything it shouldn't or done anything crazy or unrelated to what I asked.
185.
by ilaksh 5mo ago
Which of these alternatives actually doesn't use Google under the hood?
186.
by ilaksh 5mo ago
Given the intense pressure to find more efficiency, 5 years is probably enough to see deployment of radical new compute-in-memory paradigms such as nitrides-based ferroelectric devices that have a few orders of magnitude more efficiency. I
187.
by ilaksh 5mo ago
In less than five years it will be clear that unaugmented humans are a bottleneck. So many jobs as such may be limited to humans with high bandwidth BCI connected to AI, that effectively gives them an AI exobrain. But even then the AI would
188.
by ilaksh 5mo ago
Good point to a degree, but it applies to all work, not just software engineering, and even before they get into the workforce. Young people are not going to have any depth of skills or knowledge if they literally just act as a go-between f
189.
by ilaksh 5mo ago
The DHS has made many communications that were openly white supremacist. It's not just an unfair situation with legal technicalities. Their views and plans are more extreme and dangerous than our society is able to accept as reality, s
190.
by ilaksh 5mo ago
What percentage of these are variations of the good old fashioned buffer overflow that would be impossible with Rust?
191.
by ilaksh 5mo ago
Surely some significant fraction of the billions and billions are being spent on intensive R&D to for example increase scale of MRAM modules or radical new paradigms like nitrides-based ferroelectric compute-in-memory? I think there&#x2
192.
by ilaksh 5mo ago
I'm an atheist, but most of what I have heard from popes in recent years seems like sound and possibly needed advice. Also, even though I feel AI and robotics are very important for progressing humanity, I think that much of the world
193.
by ilaksh 5mo ago
Amazing. Is it possible to do this with Qwen 3.6 27B? Will it work with quants (I assume so)?
194.
by ilaksh 5mo ago
I want something like this but not only for my own computer but also for client projects or stuff I might run in cloud GPUs. Because the core idea of having a strong model that is efficient and doesn't require a cluster still applies t
195.
by ilaksh 5mo ago
Yes I assume some kind of LLM system will eventually make its way into OS package registries and eventually there will be a standard of some sort. Already OpenAI's API is the de facto standard available from all providers. But who know
196.
by ilaksh 5mo ago
It does just barely work now after they spent billions, and they may still fall back to cloud LLMs for a significant number of things. This is a way that everyone can get that on the actual Apple Watch or local phone for any application the
197.
by ilaksh 5mo ago
Good idea. Could you make that.
198.
by ilaksh 5mo ago
I think GLM 5.1 or Kimi 2.6 could substitute for this type of purpose.
199.
by ilaksh 5mo ago
Oh gotcha. Fixed my comment.
200.
by ilaksh 5mo ago
Hmm.. this might make it feasible to build something like a command line program where you can optionally just specify the arguments in natural language. Although I know people will object to including an extra 14 MB and the computation for
201.
by ilaksh 5mo ago
One aspect of this may be Chinese laws, which I am assuming if they don't already require the ability to monitor or even censor what gets printed, they will soon. Even in the US we are starting to have legislation related to blocking f
202.
by ilaksh 5mo ago
Again that is a problem with the group of people and how they use technology rather than the technology itself. I will die on this hill: AI _properly_ integrated into education will be a huge improvement for students because it will enable
203.
by ilaksh 5mo ago
> The people who rely on it are not cool. That's the only statement that's true. Admiting to AI use is unfashionable in the western world at this time. But how much would you like to bet that 90% of those students who were booi
204.
by ilaksh 5mo ago
He says he went several months without having to do a code review and it worked the vast majority of the time. That's incredibly impressive work by the AI. AI may default to mediocre and often somewhat buggy code unless you iterate bec
205.
by ilaksh 5mo ago
There have been and are many, many similar no code or low code tools. And actually Visual Basic 6 still works. What changed was the level of popularity and acceptance of those tools. I think it comes down to trends and programmers psycholog
206.
by ilaksh 5mo ago
I understand that people have had bad experiences and in general I don't want to pay for software components either. But it's weird how much hate there is against developers sometimes just for even trying to make money from their
207.
by ilaksh 5mo ago
I was hoping that the visuals going through the BCI also was implied since I didn't indicate any other new technology for holograms.
208.
by ilaksh 5mo ago
30 years down the line a human will wake up in his climate controlled bed in an idyllic large scale people-zoo, think about what information he wants, and immediately his 900TB ferroelectric compute-in-memory exobrain will read his thoughts
209.
by ilaksh 5mo ago
No, AI capabilities of some sort are obviously important. But I know a lot of people don't appreciate that. But you aren't seriously suggesting that graphics hardware is irrelevant are you?
210.
by ilaksh 5mo ago
Says nothing about AI capability or even graphics. I am skeptical about the value.

More ›