HN user

burtonator

5,915 karma
Posts86
Comments923
View on HN
news.ycombinator.com 2y ago

Ask HN: Why don't VCs just "suck it up" and pay founders a competitive salary?

burtonator
44pts71
news.ycombinator.com 3y ago

Ask HN: Best design for an Alexa-like voice powered ChatGPT

burtonator
1pts0
news.ycombinator.com 4y ago

Ask HN: What do you think of GraphQL?

burtonator
14pts9
news.ycombinator.com 4y ago

Ask HN: What do you do and what's your consulting rate?

burtonator
98pts109
github.com 5y ago

Google Cloud is blocking Firebox and Safari extensions

burtonator
2pts0
news.ycombinator.com 5y ago

Ask HN: How would a reverse captcha work? A challenge that filtered out humans

burtonator
6pts4
news.ycombinator.com 5y ago

Ash HN: Best practices for key management /w web-based client-side crypto?

burtonator
1pts1
news.ycombinator.com 5y ago

Ask HN: What are some good end-to-end crypto frameworks for JavaScript/Browsers?

burtonator
2pts0
news.ycombinator.com 5y ago

Ask HN: How does your startup estimate long-term releases/roadmap using agile?

burtonator
6pts0
news.ycombinator.com 6y ago

Ask HN: Why don't browsers have billing+shipping APIs? This would kill Amazon

burtonator
26pts13
getpolarized.io 6y ago

Bored due to quarantine? Here are the top PDFs posted to Hacker News in 2019

burtonator
3pts0
getpolarized.io 6y ago

Building a Citation Manager for PDFs on Google Cloud (Firebase)

burtonator
8pts1
news.ycombinator.com 6y ago

Ask HN: Using shock images like Goatse to fight online scammers

burtonator
2pts3
getpolarized.io 6y ago

Show HN: Top PDFs Posted to Hacker News in 2019 Computed from Internet Archive

burtonator
21pts9
getpolarized.io 6y ago

Bill Gates Reads by Writing (In the Margins)

burtonator
11pts3
getpolarized.io 6y ago

Over 500 Top PDFs Posted to Hacker News in 2019

burtonator
4pts0
news.ycombinator.com 6y ago

Ask HN: Service to hire a temporary pair programmer

burtonator
9pts4
getpolarized.io 6y ago

Dark Mode by Default – 95% of People Prefer Dark over Light Mode

burtonator
241pts222
news.ycombinator.com 6y ago

Ask HN: Pairing two users to resolve the 'bystander' effect of donations

burtonator
1pts0
getpolarized.io 6y ago

What's Next for Offline Web Capture of Web Pages in Polar

burtonator
1pts0
getpolarized.io 6y ago

Roger Craig Used Polar and Anki to Win Jeopardy

burtonator
3pts1
www.datastreamer.io 6y ago

Language Classifying 5TB of Web Content per Day

burtonator
2pts0
www.inputneuron.io 6y ago

I'll be the cofounder of your $1B startup.First,send me $200 for chocolates (ha)

burtonator
2pts0
news.ycombinator.com 6y ago

Ask HN: A 3rd party service to validate Twitter/Facebook accounts are humans

burtonator
1pts0
getpolarized.io 6y ago

Minimizing Defection Might Be the Solution to the Tragedy of the Commons

burtonator
21pts47
getpolarized.io 6y ago

Minimizing Defection Might Be the Solution to the Tragedy of the Commons

burtonator
2pts0
news.ycombinator.com 7y ago

Ask HN: What's a technology that died that you wished was still here

burtonator
14pts26
getpolarized.io 7y ago

GPT-2, the Destruction of the Web and Artificially Intelligent Textbooks

burtonator
3pts0
news.ycombinator.com 7y ago

Ask HN: A browser that clicks on every ad as a protest against Google?

burtonator
21pts27
news.ycombinator.com 7y ago

Ask HN: Have any OSS projects asked users to donate to build new features?

burtonator
1pts0

I'm honestly confused as to why it is doing this and why it thinks I'm right when I tell it that it is incorrect.

I've tried asking it factual information, and it asserts that it's incorrect but it will definitely hallucinate questions like the above.

You'd think the reasoning would nail that and most of the chain-of-thought systems I've worked on would have fixed this by asking it if the resulting answer was correct.

Prompt Caching 2 years ago

Autoregressive models can't just resume so they have to re-parse the entire prompt again for each execution.

By caching them they resume from where it left off from before thereby completely bypassing all that computation.

For large contexts this could save a ton of compute!

I think this feature and structured outputs are some of the biggest inventions in LLMs this year.

If you're ever in a situation similar to this, run as fast and as far as you can.

I really really want to underscore this point.

You're literally standing on top of ground and under that is boiling water.

If that breaks and you fall in you're going to be in boiling water with no way to get out and you will die screaming.

Also NEVER walk on ground that has no vegetation. If you look around a geyser you will see that the ground is white and has no vegetation. That's because the temperature is too high and it has water under it that's heating the ground.

Walk on that and there's a chance you will fall in.

In the back country there are no fences so you can fall right through the crust.

He got us out of Afghanistan too... this was a huge win and was very risky. Doesn't get enough credit for this one.

Agreed on NATO + Ukraine too but I wish they would allow us to strike Russia directly even with NATO equipment.

There must be a relationship here between Shannon's estimation that english is 1 bit of entropy per character and is highly redundant and 'easy' to predict.

A highly advanced AI could compress the text and predict the next sequences easily.

This seems like a direct connection like electricity and magnetism.

And maybe that's why English needs to be about 1 bit because we're not very intelligent.

In the 2000s I was addicted to Elisp and contributed a ton of OSS code including JDE, EDEE, and tons of other tools.

But... I had just a MASSIVE amount of code that was literally just for me.

Emacs basically became my OS.

Emacs allowed you to just eval code on the fly and the IDE would just adapt. No reload required. So if you wanted to do stupid stuff like make control+enter open the current URL at the cursor, you just write a three line script. Then you add it to your elisp on load.

... but mine got WAY out of hand. It was just mountains of code.

That's not really what's happening here I think.

I think he's just not sure what's ok or not ok to write about. Nothing he wrote here was problematic. No secrets and just sort of mild criticism.

The next major leap in LLMs (in the next year) is probably going to be the prompt context size. Right now we have 2k, 4k, 8k ... but OpenAI also has a 32k model that they're not really giving access to unfortunately.

The 8k model is nice but it's GPT4 so it's slow.

I think the thing that you're missing is that zero shot learning is VERY hard but anything > GPT3 is actually pretty good once you give it some real world examples.

I think prompt engineering is going to be here for a while just because, on a lot of task, examples are needed.

Doesn't mean it needs to be a herculean effort of course. Just that you need to come up with some concrete examples.

This is going to be ESPECIALLY true with Open Source LLMs that aren't anywhere near as sophisticated as GPT4.

In fact, I think there's a huge opportunity to use GPT4 to train the prompts of smaller models, come up with more examples, and help improve their precision/recall without massive prompt engineering efforts.

I started working on my own version of this but what I found is that the text extraction part is the key.

I looked at using MathPix but it has images as part of its output.

That would be fine, but I don't have the GPT4 version with image support.

It's not GA yet.

Heck. Ignoring speed, it would probably just be easier to have GPT4 index the raw images.

... so the details matter here.

We're stuck here for a while due to the size, and cost, of the larger models.

The main reason I want a non-cloud LLM is that I want one that's unaligned.

I know I'm not a criminal and I want to stop being reprimanded by GPT4.

What I'm most interested here is fine tuning the model with my own content.

That could be super valuable especially if we could get it to fact check itself, which you could with a vector database.

Honestly, I think it's completely unfair for AIs to train on this data.

I work in ML so I'm aware of the consequences but society wasn't.

My step-daughter is finally crushing it as an graphics artist and she is really pissed at tools like Midjourney.

I asked her about it and she said "yes, they steal the artwork of real artists and generate fake knockoffs" ... and I don't think her opinion is invalid.

This is senseless and a huge loss.

Bob was a great guy.

I met him briefly when he worked at Google. He was just starting to work on Guice and I was skeptical of dependency injection and we talked about it for an hour.

I went home and did a heads down and Guice was a major impact on my coding for the next ten years.

I bumped into "Crazy Bob"a few more times and just an insanely nice guy.

He was also murdered at Main and Folsom right in downtown SOMA in SF. It was 2 blocks from my former apartment.

SF is just almost not worth it at this point.