dbish

12 ms·

151.
by dbish 2y ago
really? i tend to look to them for ideas
152.
by dbish 2y ago
they also seem to have shifted to hardware/"AI" wearables (lavalier mics?). hardware is not a promising shift imho
153.
by dbish 2y ago
We've done a bit of both for our screen seachable loom-like screen recorder, the problem is that the accessibility APIs differ greatly between Mac and Windows if you want to be OS agnostic and even on Windows all the apps tend to do th
154.
Snippy: An OSS Rust-Based AI Assistant with Eyes(timdbg.com)
4 points by dbish 2y ago|0 comments
155.
by dbish 2y ago
It really doesn’t. It has no history or proactive understanding, it only runs a query when you ask it
156.
by dbish 3y ago
My friend’s kid really likes to email people so I thought it would be neat to build a little auto-responder powered by gpt with different personalities (prompts), got this up and running using flask + gpt assistants api that uses assistants
157.
by dbish 3y ago
Neat, are you able to handle threads/responses well even if the responder is not from gmail? I worked on a toy project and put together a library for myself over time that was harder to do then I expected due to the various email nuanc
158.
by dbish 3y ago
Does this work well on screenshots of a desktop or webpage? We’ve found a lot of the document segmentation/layout parsing models and tools are very tuned to really paper docs only, the kind of thing you’d see in a research paper or sha
159.
Show HN: GraggNet – Human Surrogates for AI Agents (April Fools')(augmend.com)
2 points by dbish 3y ago|0 comments
160.
by dbish 3y ago
More reps/iterations to learn from
161.
by dbish 3y ago
It’s a bit different when you’re replacing or creating the head of a division since they’ll replace their own lieutenants and onwards down
162.
by dbish 3y ago
Eh I think some startup culture injected into Microsoft is a good thing, whatever that might mean. Have to shake things up
163.
by dbish 3y ago
Which?
164.
by dbish 3y ago
One of the reasons I tend to use any of their options through Azure where available. Azure support has a more straight forward (though still sometimes slow) process for account issues.
165.
by dbish 3y ago
I always use screenshot based fallbacks, so the old SEO tricks won't quite work for that. You want to look at it through human eyes.
166.
by dbish 3y ago
One of the big problems with RPA is that it's very very specific, and requires less natural tool interactions that we can do with the new models (or will soon be able to do). It should be as simple as having an AI system "look ove
167.
by dbish 3y ago
SATs and ACTs are pretty close to IQ tests
168.
by dbish 3y ago
There are many (many :)), metrics for NLG and NLP comparison that can be chosen. These are not simply subjective and no way to compare
169.
by dbish 3y ago
Yup. Take a test and show you should be a good fit for a school, awesome, have to be tracked throughout a relatively boring (classroom-wise, except for programming class) couple years of high school, no thanks.
170.
by dbish 3y ago
How would you compare raw intelligence or cross schools? That kind of calibration and opportunity for smart students who had other reasons to not focus or care about high school and other remedial grades is really important.
171.
by dbish 3y ago
Lol. I should add that as an option you can toggle
172.
by dbish 3y ago
Yeah it’s pretty neat and can be improved with more context for technical terms. You do risk some hallucinations
173.
by dbish 3y ago
Is there a metric you use for this comparison? Any reason you'd expect Circleback to have context or info that's better then Zoom?
174.
by dbish 3y ago
I do this with just a prompt setting context of the meeting if we have it, explaining this is a spoken meeting between n people, and asking to correct based on the entire context, acutally works decently out of the box.
175.
by dbish 3y ago
Totally agree on the baseline. We’ve found that adding multimodal data like what was onscreen to be a big help to improve over this, which is a little more complex. Helps more to add action data to like who was typing in what, where the mou
176.
by dbish 3y ago
Visual understanding is key to anything that isn’t just talking heads in a discussion and what almost all of these tools are missing, I think because it’s a lot harder then using whisper and OpenAI (or equivalent) and a lot more expensive&#
177.
by dbish 3y ago
That’s really interesting to hear. I did a bunch of interviews with devs around real time multiplayer collab since I wanted some improvements there to things like screenheroand was surprised at how infrequently they actually did it. We focu
178.
by dbish 3y ago
Site would just not run if you're using EC2 directly and the machine can't handle it. Not sure what cloudflare does, i don't use it, just AWS directly :)
179.
by dbish 3y ago
Wild. I’ve worked with a bunch of dev teams across FAANGs and smaller companies and never had a whole team of people on Mac even with a small 5 person team.
180.
by dbish 3y ago
If you run on a single EC2 instance and you aren’t running an auto scaling cluster or anything of the sort, it would be pretty hard to get a huge bill. I much prefer that and the chance that it goes down then autoscaling or severless. Most

More ›