HNHacker News
TopNewBestAskShowJobs

heyitsguay

1,081 karma · joined January 5, 2017

submissionscomments
heyitsguay··on Deep fake video of Putin declaring martial law is broadcast in parts of Russia
Lol, so now what happens when they do actually try to declare martial law in response to border incursions?
heyitsguay··on Cargo Cult AI
And moreover, this is more than a philosophical difference. As our understandings of AI and neuroscience improve we can now talk about "world models" as something almost tangible and constructible. Complex animal nervous systems build and use embodied world models to interpret the senses and allow for tractable future prediction and action generation. Yann Lecunn's paper on AGI from August 2022 gets into defining computational equivalents, and multimodal AI models like PaLM-E are taking first steps toward explicitly tying LLM-style text prediction with collections of perceptual observations about the environment.

So it's not that artificial systems cannot, as a category, have models of their environment and the actions they can take within it, this "cargo cult AI" concept comes up when people jump the gun and see those capabilities in much simpler systems, even including ChatGPT-4. And it's disheartening to see narrative on the subject driven more and more by people who have not taken the time to learn about the subject matter, for all the interest they seem to show in talking about it.

heyitsguay··on Mojo: The usability of Python with the performance of C
You mention coming into this from the ML world, but any mention of timelines to leveraging Python's ML ecosystem seem absent. Any idea what's in the works in that regard?
heyitsguay··on Ask HN: How to find local AI meetup and events?
I help run a weekly AI meetup at NoiseBridge. You can come check us out Wednesdays at 8 or join our Discord: https://discord.gg/cBpbz9VedY .

It's a mix of things at different levels. Some is high level "here's LangChain or Automatic1111", some is more in-depth paper reads. I covered BLIP2 recently, someone else did PaLM-E two weeks ago, I'm gonna touch on SAM this week.

heyitsguay··on DINOv2: State-of-the-art computer vision models with self-supervised learning
And trainable on only 96 GPUs? Loving this recent trend in vision models that are borderline tractable to replicate.
heyitsguay··on Large Language Models Are Human-Level Prompt Engineers
I've messed with ChatGPT before but this weekend I tried to use it seriously to hack together a python demo in the computer vision space. I appreciate it and it's a way better rubber duck for me to talk my ideas through with and generate sample snippets, but hallucination is a problem, and the more intricate and customized the code being developed, the more prone it's been to misinterpretation.

I'm getting some good code starting points, and talking the idea through step by step with a chatbot is really helping me clarify what i need to do, but I've still got API docs open for the libraries it uses because it likes to make up functions, including the core one on which the project logic hinges. (But that's ok, because now I know I just need to write a function that works that way and i can do that).

Pretty cool and helpful! And I can imagine it getting better with GPT-4 or code-specific tooling. But it's generating value on the order of like.. many other SaaS offerings that have come onto the scene that try to ease pain points in coding workflows. Versus value of the sort that upends society and my entire way of life. A great new tool that I should learn about to make rote bits of my activities faster and easier, a story that's a bit more familiar in tech than some of the more breathless stories about AI make it sound.

heyitsguay··on Large Language Models Are Human-Level Prompt Engineers
For sure NLP writ large requires lower accuracy levels than AV, like how computer vision writ large has many applications that require lower accuracy levels than AV. And indeed over the past few years what we've seen play out in CV deployments in the real world is incremental gains in controlled environments built on gobs and gobs of application-specific engineering, vs breakout success steamrolling over industries with standardized turnkey solutions

I say that as someone who works in the space and loves it tbc. And I fully expect to see some wild stuff make it IRL this decade in both NLP and CV, I just think the rubber hits the road a bit more slowly than pop social media discourse would have one think.

heyitsguay··on Large Language Models Are Human-Level Prompt Engineers
LLMs are having a moment in 2023 like self driving cars were having in 2015. Some really cool demos following a lot of hard work, too much hyperbolic speculation that mass real-world job-destroying deployments are right around the corner, not enough appreciation of how few commercial applications are ok with 99% (or even 99.9%) accurate solutions. Real value being created, but still requiring lots of human ingenuity and grunt work to unlock it.

As a decent first-order metric - follow the ratio of companies getting money for using LLMs to do something, to companies getting money for providing LLMs and associated tooling to others. The bigger that ratio gets, the more real world impact LLMs are having.

heyitsguay··on Swin Transformer: Hierarchical Vision Transformer Using Shifted Windows
Something I've noticed is that there's a ton of work going into better vision transformers, but whenever there's a big new multimodal result or something, everyone's just using ViT. What gives? Like in CLIP or PaLM-E, they very easily could have used another vision architecture, it's not like they were running all modalities through the same encoder.
heyitsguay··on It takes a body to understand the world: language AIs don’t know what they say
Yeah, but build a new model that injects camera and microphone observations as well as robotic hardware internal state estimates into the same embedding space as your favorite LLM, fine tune the whole thing, and suddenly your text generator is referencing a semantic space that coreferences observed physical relationships. You break out of the circular reference loop pure LLMs suffer from. Embodiment and grounding all at once.

Google's recent PaLM-E paper shows how much progress is already being made towards exactly this sort of multimodal embedding soup. Things are going to keep getting wilder!

heyitsguay··on Simply explained: How does GPT work?
That's not what multi-head attention means. Multi-head attention is the use of learned projection operators to perform attention operations within multiple lower-dimensional subspaces of the network's embedding space, rather than a single attention operation in the full embedding space. E.g. projecting 10 512-D vectors into 80 64-D vectors, attending separately to the 8 sets of 10 embedding projections, then concatenating the results together to reform 10 512-D vector outputs.

In fact the projection operations are the only learned part of a Transformer's self-attention function -- the rest of self-attention is just a weighted sum of the input vectors, where the weights come from the (scaled) vector correlation matrix.

heyitsguay··on HuggingFace and Open Source AI Meetup in SFO Mar 31st
Is the event totally full? I'm in the Mission but I guess I missed the earlier announcements. I'll keep an eye out for future events!
heyitsguay··on The case for slowing down AI
Thank you, you said it better than I did replying to another comment, but I feel about the same. It's disheartening how many outlets we have on social media to engage with big exciting narratives that are just detached from reality, and now here's one more.
heyitsguay··on The case for slowing down AI
The problem is this kind of discourse is coming almost entirely from people who haven't made a good-faith effort to understand the technology. You're engaging with Hollywood illusions that don't resemble reality. There are genuine concerns related to AI developments but they have nothing to do with "alignment", it's all that boring stuff like shifts in power dynamics and biases in decision making. Really important, but not exciting enough for a gripping pop narrative. Less basilisk, more... failing to assemble adequately debiased training datasets for your mortgage approval app, or realizing someone can actually identify certain individuals in an anonymized medical dataset.

If you find yourself feeling anxious about AI, spend time learning how modern transformer-based predictive and generative models actually work. From resources rooted in math, statistics, and computer science; not doomer bloggers.

heyitsguay··on USCO tells judge that art created without human involvement is not protectable
Bit of a clickbait title -- this is about an "AI" producing its own artwork autonomously (aka the same as procedural art i guess? Who knows), not about artwork a human produces using tools like DALL E or Stable Diffusion.

This is consistent with recent rulings about copyright by AI outside the art domain.

heyitsguay··on Darwinian Gastronomy: Why We Use Spices (1999)
And a touch of garlic powder!
heyitsguay··on Understanding the limits of large language models
Lol, lots of people spouting off about how they imagine AI works these days. This is not an accurate description of the GPT2/3 model architectures.
heyitsguay··on S.F. apartment rents fell again
Since SF is on the narrow tip of a peninsula and surrounded by water on 3 sides, the temperature fluctuates much less than locations inland or even just along the coast. The temperature basically hovers around 60F year-round, a few degrees cooler in the winter, a few degrees warmer in the autumn.

I didn't know this before moving here, I don't know if I'd call it _amazing_ weather but it's certainly odd to watch the rest of the country deal with insane heat and cold while SF basically just... stays at 60.

heyitsguay··on The Transformer Family
For a while now, an answer I've seen is to start with "Attention Is All You Need", the original Transformers paper. It's still pretty good, but over the past year I've led a few working sessions on grokking transformer computational fundamentals and they've turned up some helpful later additions that simplify and clarify what's going on.

You can quickly get overwhelmed by the million good resources out there so I'll keep it to these three. If you have a strong CS background, they'll take you a long way:

(1) Transformers from Scratch: https://peterbloem.nl/blog/transformers

(2) Attention Is All You Need: https://proceedings.neurips.cc/paper/2017/hash/3f5ee243547de...

(3) Formal Algorithms for Transformers: https://arxiv.org/abs/2207.09238

heyitsguay··on Andromeda Invaders: A retro-style game written in HTML5, Canvas, and Web Audio
Very nice! Reminds me of a little HTML5 Asteroids clone/mutant I wrote a while ago: https://heyitsguay.github.io/asteroids/

I should figure out how to add sound. It's always inspiring to see the great work of others :)

heyitsguay··on Techniques to improve reliability
Is this something one could try to quickly implement alongside NanoGPT? Seems like a pretty straightforward, concrete idea, once you decide where you want those tokens to fit into downstream attention layer inputs. Evaluating relative performance on a small scale could give indication of if it's worth trying at larger scales, unless it's one of those things that doesn't help until your model is huge.
heyitsguay··on ChatGPT listed as author on research papers: many scientists disapprove
This is actually a case of credulous laypeople drawing attention to something academics know enough to dismiss (and to communicate to others that it is dismissal-worthy)
heyitsguay··on Microsoft’s VALL-E that can clone voice from a 3-SEC audio clip
It raises awareness of the fact that such techniques are possible.

The generative AI results of the past year straddle the invention/discovery divide -- I've seen snarky internet takes recently along the lines of "why are tech bros solving art and poetry instead of our tedious labor", but what people not clued into the field don't get is that:

(a) This generative stuff is easier specifically because it doesn't have to interact with the real world. It's just data manipulation, there's a ton of it to learn from, and it's basically ok if a lady in a picture has six fingers or if the poem you generated has an incorrect meter on line 8. IRL interactions are harder to get data on, and nobody wants their bulldozer AI to hallucinate that the building next door is part of what needs to be demolished.

(b) These generative tools can exist precisely because there is regular statistical structure in (art/language/music). It takes some work to capture it, but it's the information's structure itself that enables, e.g., cloning a voice from a few seconds of recordings. Making research that exposes that fact taboo just means that actors who exploit it anyway will have a bigger advantage in abusing it.

It can also help to understand where a line of research comes from -- in VALL-E's case, it's an outgrowth of research into neural audio compression, an area with very clear advantages for information technology. If one starts shipping that technology and parts of it can be used for few-shot cloning of others' voices, it seems better that we're all aware of the fact.

heyitsguay··on ChatGPT integrated in a game for NPC dialogue interactions
Pretty cool, but it seems like this wouldn't be as exciting as the demo makes it seem for a real game until LLM grounding is better solved. Until then, the dialogue can only be fancy approximately-correct window dressing on top of the game engine which isn't open-ended enough to make use of the LLM's inventions. Instead of your NPC saying "I heard there was a dragon in Whiterun" over and over again, you can have "I heard there was a dragon in Whiterun", "Did you hear about the dragon in Whiterun?", "Someone told me they were in Whiterun and spotted a dragon", "I heard nobody saw a dragon in Whiterun[sic]" etc. Cool for what it is, but not much of a game changer until the generated dialogue can actually change the game.
heyitsguay··on ChatGPT Considered Harmful: Our Whole Society Is in Danger
I'd like to comment on this, but the author has not explicitly consented to me learning from the text.
heyitsguay··on All in one OSINT tool with web interface
Your README doesn't really seem to give a clear idea of what this tool is or why anyone would want to use it. You say to check the wiki but there doesn't appear to be much content.

In general this sort of communication seems to be a sticking point for the engineering-minded -- it's cool you built a backend in Go, but a backend for what? What's the hook that will draw me in to investing the time to check out your particular tool?

heyitsguay··on AI-powered lawyer will be first of its kind to represent defendant in court
While a pure ChatGPT-style bot would be prone to hallucinations and overconfidence, isn't law actually a fertile field for working out grounded chat LLMS?
heyitsguay··on I've just started mixing shaders with Pygame
Can you share anything about your stack or process? This looks beautiful!
heyitsguay··on Do simpler machine learning models exist and how can we find them?
Interesting, any material you can share?
heyitsguay··on Do simpler machine learning models exist and how can we find them?
As someone in the ML industry, who knows many people in the ML industry, we all know this. It's non-technical fundraisers that spread the hype, and non-technical laypeople that buy into it. Meanwhile, the folks building things and solving problems plug right along, aware of where limitations are and aren't.
← PreviousPage 5 of 15Next →