drusepth

32 ms·

151.
by drusepth 2y ago
I started incorporating LLMs into my workflows around the time gpt-3 came out. By comparison to its performance at that point, it sure feels like my junior is starting to become a senior.
152.
by drusepth 2y ago
Yeah, my best guess is they're probably including the previous N frames as context into generating the next model. This works to preserve continuity over a short amount of time (as you say, briefly looking at the ground and then back u
153.
by drusepth 2y ago
We generate 2d art, 3d models, some other in-world data, and soon animations at our game studio. The huge problems of generating pure pixels without a persistent game state behind it is illustrated well by the recent MinecraftAI craze (wher
154.
by drusepth 2y ago
The company was formed as Niantic Labs in 2010 as an internal startup within Google, founded by the then-head of Google's Geo Division (Google Maps, Google Earth, and Google Street View). It became an independent entity in October 2015
155.
by drusepth 2y ago
> It is rather about improving some vision models to predict how the backside of a building looks, right? This, yes, based on how the backsides of similar buildings have looked in other learned areas. But the other missing piece of what
156.
by drusepth 2y ago
I potentially like the concept, but it's kind of hard to take serious when you open the page and are immediately greeted by a large-busted anime girl (and an animated ticker spanning the bottom of the page, full of zoomed-in crops of s
157.
by drusepth 2y ago
"Popular" tweets (of which these bot accounts often fall into, because they're propped up by bot responses and engagement farming) are pushed into your feed even if you're not following (or engaging) with them.
158.
by drusepth 2y ago
What makes you think the FBI hates Trump?
159.
by drusepth 2y ago
> I've always dreamed of full blown sandbox games with extremely advanced NPCs (which the current LLMs can already kinda emulate) The future of gaming is going to get weird fast with all this new tech, and there are a lot of new mec
160.
by drusepth 2y ago
I'm glad there are different kinds of podcasts for different people now. I've always absolutely hated the focus on the individual humans and their personalities behind the podcast, and wished they'd be a better source of well
161.
by drusepth 2y ago
I'm glad that Kagi (and others) exist as an alternative for people who don't want generative AI in their searches. Personally, I'm excited about more generative AI being added to my search results, and I'll probably swit
162.
by drusepth 2y ago
I hate podcasts because they're so often focused on the speakers' personalities and windy, undirected things. I've tried to listen to so many podcasts and always dip an episode or two in because they devolve into people just
163.
by drusepth 2y ago
If this is the case, I imagine it'd be pretty easy to spoof your UA and see the demo, even from Chromium.
164.
by drusepth 2y ago
Scrapers have always been notoriously brittle and prone to breaking completely when pages make even the smallest of structural changes. Scraping with LLMs bypasses that pitfall because it's more of a summarization task on the whole doc
165.
by drusepth 2y ago
I'd say scrapers have always been popular, but I imagine they're even more popular nowadays with all the tools (AI but also non-AI) readily available to do cool stuff on a lot of data.
166.
by drusepth 2y ago
To be fair though, you typically end up doing that process even when you remain within the OpenAI ecosystem, because they're putting out new (cheaper) models. For example, a couple weeks ago I migrated a few game mechanics from GPT-4-T
167.
by drusepth 2y ago
Do you have any ramps/rails/etc for your desk? That sounds like such a good way to give your hands a nice break between code sessions.
168.
by drusepth 2y ago
I and a friend are working on Elixir Emporium [0], a single-player crafting RPG that uses layers of generative AI to create and inject new game data as you play. It started out as just a prototype asking what happens if you task an LLM with
169.
by drusepth 2y ago
I posted "How to Speed Up Your Computer Using Google Drive as Extra RAM" [0] ~10 years ago and have had hundreds of people over the years ask if I used Markov chains to generate the funny bits (I didn't). I've had 0 peop
170.
by drusepth 2y ago
OT: Does everyone on HN have a WaPo subscription or something? I don't think I've ever been able to actually read any article of theirs without a paywall, ever. Not even the first sentence -- literally just a title and a paywall.
171.
by drusepth 2y ago
I couldn't meaningfully do any of the three with a frozen burrito, unlike with bots and/or humans. More power to you if you can, though.
172.
by drusepth 2y ago
> Most human beings, I surmise, value human interaction more than we value object interaction. I think the key here is that you never see a human (or object) when interacting with others on the internet; your brain fills in the gaps to s
173.
by drusepth 2y ago
I do all three of these with bots (and also humans).
174.
by drusepth 2y ago
I think there's two likely scenarios here: 1) My comment is only ever going to be read by computer programs and I'll probably get a response back from a computer program (see: ChatGPT/GPTs, character.ai, AI Studio, Replika, t
175.
by drusepth 2y ago
> What is the real value of social media when most of the content posted is created by a bot? Besides the maybe 5% or so of people I interact with online that I actually know IRL, I don't really care if I'm interacting with a b
176.
by drusepth 2y ago
Could be cool. It's a nicer, smaller form factor than most of the competition today, but I'm wary of the low resolution, lack of camera, lack of speakers, and awkward bumps on the earpieces. But a 1.5 day battery (if true) sounds
177.
by drusepth 2y ago
We slashed police funding in Portland and decimated the force available (to the point where the 911 response time to a robbery in my building was ~6 hours, local businesses my friends work at have private security for employees to call inst
178.
by drusepth 2y ago
Why is Obama (I assume Michelle?) in those odds (and, interestingly, at #3 right behind Kamala)? Has she ever indicated she might run?
179.
by drusepth 2y ago
I don't know the quote you're referring to, but sexual assault by siblings is at least 19x more common (2.3%+ of children) than by adult family members (0.12%+ of children), according to the U.S. Department of Health and Human Ser
180.
by drusepth 2y ago
Downvotes also might suggest one disagrees with the opinions mistakenly (and, arguably, dangerously) presented as fact here.

More ›