HNHacker News
TopNewBestAskShowJobs

shwaj

479 karma · joined April 28, 2023

submissionscomments
shwaj··on iPhone Pocket
lol! What corner is that?
shwaj··on Redmond, WA, turns off Flock Safety cameras after ICE arrests
You missed the point: they were referring to the sentiment that lead to Trump’s election. Many of those voters , I would guess, feel that the way Trump is doing it is cruel and chaotic.

This obviously doesn’t imply that those who voted for Trump on this basis want to go back to the open border Biden days.

shwaj··on Ruby and Its Neighbors: Smalltalk
I used to find it amusing that you could say `Smalltalk := nil` in the early days of Squeak, and poof! all the classes in your system are garbage collected.

(I’m not sure that’s exactly what happened, probably the system crashed before garbage collection could happen. But it was definitely a guaranteed insta-crash).

shwaj··on BlackRock's Larry Fink: "Tokenization", Digital IDs, & Social Credit
Maybe Klaus shouldn’t have said it then.
shwaj··on Backpropagation is a leaky abstraction (2016)
Hear hear, one of my favorite comments recently.

Can’t agree more about the technical points (category error etc), and then the unexpected switch to the value of receiving constructive criticism as a gift not an attack.

Myself, I’m definitely conditioned to receive it as an attack. I’m trying to break this habit. This morning I gave some extensive feedback to some friends who have a startup. The whole time I was writing it, I was stressing out that they’d feel attacked, because that’s how I might take similar criticism.

How was it actually received? A mix I think. Some people explicitly received it as a gift, and others I’m not so sure.

shwaj··on My Impressions of the MacBook Pro M4
The part about noticing web pages loading (at most) 8ms faster due to the display is total nonsense. Many can notice the difference between 60 and 120Hz when scrolling, but definitely not for a page load. That’s less than 1/10th of the blink of an eye.

If page load seems noticeably faster, it’s far more likely that it’s simply a faster machine. Or imaginary.

shwaj··on Denmark reportedly withdraws Chat Control proposal following controversy
Nobody’s talking about a blood oath to promise never to revisit the issue. But there’s a different between leaving the door open to future reconsideration, versus pushing consistently against the wishes of the public and only backing off temporarily for tactical reasons.

And for some reason, once these things pass, it’s a one way door. When does the US public get a chance to reconsider the Patriot Act?

shwaj··on Extropic is building thermodynamic computing hardware
I think what they meant was:

Two linear transformations compose into a single linear transformation. If you have y = W2(W1*x) = (W2*W1)*x = W*x where W = W2*W1, you've just done one matrix multiply instead of two. The composition of linear functions is linear.

The ReLU breaks this because it's nonlinear: ReLU(W1*x) can't be rewritten as some W*x, so W2(ReLU(W1*x)) can't collapse either.

Without nonlinearities like ReLU, many layers of a neural network could be collapsed into a single matrix multiplication. This inherently limits the function approximation that it can do, because linear functions are not very good at approximating nonlinear functions. And there are many nonlinearities involved in modeling speech, video, etc.

shwaj··on Grokipedia and the coup against reality
Sorry, for which crimes? I know the justice system is flawed, but ideally shouldn’t there be specific laws that are broken, to justify imprisonment?
shwaj··on Keep Android Open
Yes, I recall HN commenters of the time predicting this exact outcome as a result.
shwaj··on A definition of AGI
I think you meant to respond to @noduerme, not @qudat. If so, I agree with both of you, otherwise I’m confused by what you’re saying.
shwaj··on Linux Capabilities Revisited
Fuchsia pervasively uses capabilities.
shwaj··on Willow quantum chip demonstrates verifiable quantum advantage on hardware
I’m not sure there’s a better alternative.
shwaj··on AI is making us work more
What does this mean?
shwaj··on AI is making us work more
They are obviously not saying that they think this is going to happen. It is explicitly an “if-then” thought experiment: “if technology is deflationary, then avoid civil unrest”. The tone is rightly skeptical.
shwaj··on The scariest "user support" email I've received
That doesn’t logically follow. It got this very straightforward thing correct; that doesn’t prove their response was cynical. It sounds like they know what they’re talking about.

A couple of times per month I give Gemini a try at work, and it is good at some things and bad at others. If there is a confusing compiler error, it will usually point me in the right direction faster than I could figure it out myself.

However, when it tries to debug a complex problem it jumps to conclusion after conclusion “a-ha now I DEFINTELY understand the problem”. Sometimes it has an OK idea (worth checking out, but not conclusive yet), and sometimes it has very bad ideas. Most times, after I humor it by gathering further info that debunks its hypotheses, it gives up.

shwaj··on Abandoned land drives dangerous heat in Houston, study finds
If he’s that drunk all the time, the uncle might want to wear a helmet.
shwaj··on The Unix Executable as a Smalltalk Method [pdf]
Yes it was Alan Kay. https://wiki.c2.com/?AlanKayOnMessaging

Another relevant page on the same site is: https://wiki.c2.com/?AlanKaysDefinitionOfObjectOriented

shwaj··on The Unix Executable as a Smalltalk Method [pdf]
Not really, because you still recompile and start the program from scratch, rather than modifying the code that executes on the still-existing data structures.

Edit: rather than just naysaying, it occurred to me to reference the notion of Orthogonal Persistence, which the image-based approach provides (not without drawbacks) but IDEs typically don’t. Previous HN discussion: https://news.ycombinator.com/item?id=39615228

shwaj··on Apple M5 chip
How does it “jam up the airwaves” if its operating at a different frequency than the devices you say it will be jamming?
shwaj··on Intel Announces Inference-Optimized Xe3P Graphics Card with 160GB VRAM
Oops, sorry I misread. That makes more sense in context.

Yeah, I remember all the hype about the first Nvidia chip that offloaded “T&L” from the CPU.

shwaj··on Intel Announces Inference-Optimized Xe3P Graphics Card with 160GB VRAM
I think they’re using “vector” in the linear algebra sense, e.g. multiplying a matrix and a vector produces a different vector.

Not, as I assume you mean, vector graphics like SVG, and renderers like Skia.

shwaj··on Why is everything so scalable?
I recall hearing, more than a decade ago IIRC, that facebook.com ran as a monolith for much longer than you would have expected. Perhaps someone with direct experience can comment?
shwaj··on Meta Superintelligence Labs' first paper is about RAG
At some point it becomes difficult to improve the O() complexity. How do you do better that the O(n-squared) of the Transformer, with acceptable tradeoffs? Many big brains in all the big labs are very aware of the importance of algorithmic advances. There is no low hanging fruit, but they're doing their best.

Then in parallel to that looking at compiler optimizations, and other higher-level algorithmic innovations such as Flash Attention (a classic at this point) which had a drastic impact on performance due to cache awareness, without changing the O() complexity.

shwaj··on A small number of samples can poison LLMs of any size
I’m not a practitioner. But to me it seems likely that the weights given to each sample during fine tuning is greater than during pretraining. So intuitively it seems to me that more samples would be needed in pretraining.
shwaj··on AI models need a virtual machine
Fuchsia seems like a potentially practical OS for constraining the operations of AI models. As an object capability operating system, each component (and hence the process that one is instantiated in) has access only to the capabilities that it is explicitly granted.
shwaj··on Measuring the environmental impact of AI inference
Are you sure? It wouldn’t shock me, but they specifically say “Gemini Apps”. I wasn’t familiar with the term, but a web search indicated that it has a specific meaning, and it doesn’t seem to me like web search AI summaries would be covered by it. Am I missing something?
shwaj··on The Rise and Fall of Music Ringtones: A Statistical Analysis
> Today, I consider anyone who isn’t an emergency responder with a phone that makes noise a selfish asshole.

You’re probably getting a lot of false positives. Is it ok to answer a vibrating phone around you, or do I need to leave the vicinity first?

shwaj··on What Can a Cell Remember?
You might want to look into the work of Michael Levin’s lab to give you a broader perspective on the intelligence of individual cells and tissues. Your machine metaphor is arguably as misleading as anthropomorphization.

Edit: did you read the article? The examples described go far beyond your straw man about undiscovered receptors.

shwaj··on François Chollet: The Arc Prize and How We Get to AGI [video]
In this context I believe it refers to models that are trained to generate an internal dialogue that is then fed back in as additional input. This cycle might be performed several times before generating the final output text.

This is in contrast to the way that GPT-2/3/“original 4” work, which is by repeatedly generating the next finalized token based on the full dialogue thus far.

← PreviousPage 5 of 10Next →