itissid

20 ms·

241.
by itissid 2y ago
Sounds a bit more like you want to do something reranking-ish. Ideally, you would train a retrieval system to retrieve the most relevant pages which would inturn have been trained on a dataset not very different from MS-Marco. This would ge
242.
by itissid 2y ago
How Do we know that gpt-4o.has not been trained on this dataset?
243.
by itissid 2y ago
Since it says on the blog that its only images, text and audio input, does GPT-4o likely have a YOLO like model on the phone to pre-process the video frames and send BBoxes to the server?
244.
by itissid 2y ago
Have you tried replacing the input string with a random but fixed mapping and obfuscate that its chess(like replace the word 'chess' with say, 'an alien ritual practice') and see how it does?
245.
by itissid 2y ago
Noob here. Can one make the following deduction: In transformer based architectures, where one typically uses variation of attention mechanism to model interactions, even if one does not consider the autoregressive assumption of the domain&
246.
by itissid 3y ago
I recall first haring and reading about soft robot application from veritasium: https://www.youtube.com/watch?v=058hRtaCWC0 I think you must have heard of it? In the realm of rescuing or assisting people unreachable otherwi
247.
by itissid 3y ago
I recall paragraph2Vec. It was the earliest way to experiment with position embeddings that I recall reading. At first when I read it, it felt kind of crazy/weird that it worked: You feed in the paragraph ID(yes just an integer of the
248.
by itissid 3y ago
Would that also throw projects like https://en.wikipedia.org/wiki/Large_Interferometer_For_Exopl... also into question?
249.
by itissid 3y ago
> Much of the meaning of text is hidden between the lines: without understanding why statements appear in a document, a reader has only a shallow understanding. This seems not true when I and most people I know read things. I would argue
250.
by itissid 3y ago
An 8bit microcontroller https://www.cpu-world.com/CPUs/Z8/Zilog-Z8681PS.html The rom is from AMD https://www.ebay.com/itm/303174798174 I can spot a DAC unit from Motorola too
251.
by itissid 3y ago
Every one should go through this rite of passage work and get to the "Attention is all you need" implementation. It's a world where engineering and the academic papers are very close and reproducible and a must for you to p
252.
by itissid 3y ago
And we have been told that with innovation and disruption, a new breed of jobs and skill sets are created. But we don't know (or are very bad at predicting) what those would be, especially now that the world has 100's millions of
253.
by itissid 3y ago
So what happens to the film industry now? - Local/Bespoke high quality video content creation by ordinary Joes: Check. - Ordinary joes making fake porn videos for money: Check. - Reduce cost for real movies dramatically by editing in A
254.
by itissid 3y ago
How does one cope with this? - Disruptions like this happen to every industry every now and then. Just not on the level of "Communicating with people with words, and pictures". Anduril and SpaceX disrupted defense contractors and
255.
by itissid 3y ago
I've to go lie down...
256.
by itissid 3y ago
GPTs only generate specific answers if they are trained using RHLF to prefer certain answers to others. Won't this mean that coming up with a contract that meets a special individual's case will require that much more fine-tuning?
257.
by itissid 3y ago
Maybe the most amazing and ultimate end-game here would be an API trust layer that can separate the most private data from 3p APIs. - Location anonymization to fetch any resources like restaurants. - Sensitive personal data encryption to
258.
by itissid 3y ago
Yeah exactly right. If the growth > 0 you can continue borrowing and keep interest rates low.
259.
by itissid 3y ago
Great teachers(if you can find them): Andrej Karpathy, Andrew Gelman & Ben Goodrich(Columbia), Subbarao Khambampati(ASU) to name a few I know of. Go Where the hard problems are (or find someone who is doing it): If you don't have a
260.
by itissid 3y ago
Protip here: Do it by hand, though when you start getting tired of typing latex you can switch to IDE and let github copilot complete it(it will mostly be incorrect) and then you can go in and fix its mistakes, it still saves a bunch of tim
261.
by itissid 3y ago
Or do Andrej Karpathy's youtube backprop ninja exercise on jupyter notebook by hand no?
262.
by itissid 3y ago
You know what would be fun(and awesome), if the thing just let you download a "master" Massive Movie Model(MMM) and send in smaller size prompt data to generate any of the movies in that catalog on demand. It would shift all that
263.
by itissid 3y ago
https://www.reddit.com/r/LocalLLaMA/comments/17kcgjv/how_doe... reddit thread has some ideas. What I can draw from reading of that thread is that you can buy a Desktop Rig with 200GB memory bandwidth (co
264.
by itissid 3y ago
https://www.reddit.com/r/LocalLLaMA/comments/17kcgjv/how_doe... reddit thread talks about some of the pros and cons of a m3 max with 128 Gb costing ~5-6K
265.
Ask HN: How to subscribe for specific repeated stories on HN
2 points by itissid 3y ago|2 comments
266.
by itissid 3y ago
Apparently this guy like a bunch of others like https://github.com/ggerganov/ggml are implementing transformers from papers for people that want them. Pretty cool.
267.
by itissid 3y ago
Like in Workflow Guided explorations like: https://arxiv.org/pdf/1802.08802.pdf , for example, if you said "Forward the email" the Bot needs to sample from a set of actions which in this case would be DOM elem
268.
by itissid 3y ago
Just to be clear the "scalable" part here is assuming the dataset of human demonstrations is available it can in a closed loop learn better than before? In other words, am I right in understanding that the mapping from Abstract
269.
by itissid 3y ago
Could one of the roomba companies now make those things recognize the objects and avoid them, instead of dead reckoning or RF based obstacle avoidance?
270.
by itissid 3y ago
With cost of storage pretty low and AI being so good, I wonder if having a home based thing for everyday use, especially photos (Sort, tag, Memory creation, junk detection etc.) will become easy enough that cloud usage declines. There are s

More ›