- tornato7 parentTaxis only got that way because of the rules that created an artificial scarcity / monopoly on Taxi rides. Uber flouted those rules, and it was much better than a Taxi, so their lawyers cleaned up the mess later.
- I would like it noted that I had Jorge Hirsch as a professor many years ago, and his grading was extremely harsh - you needed to be flawless at showing your work, not just getting the right answer. I remember getting a 14% on one of his tests, the worst score I ever got on any test in my life. It turned out to be a B+ after the curve was applied.
Hence, it doesn't surprise me at all to see that he was the one to call out Dias. Some things never change!
- Thanks. This is about on par with what Gemini Ultra responds, whereas GPT-4 responds better (if oddly phrased in this run):
> The bunny has fur on its hind feet that resembles a pair of white shoes. However, one of the front paws also has a patch of white fur, which creates the appearance that the bunny has three "white shoes" with one "shoe" missing — hence the circle around the paw without white fur. The humor lies in the fact that the bunny naturally has this fur pattern that whimsically resembles shoes, and the caption plays into this illusion by suggesting that the bunny has misplaced one of its "shoes".
- This is my highly advanced test image for vision understanding. Only GPT-4 gets it right some of the time - even Gemini Ultra fails consistently. Can someone who has access try it out with Opus? Just upload the image and say "explain the joke."
- Funny because today I find the install process for Mac much simpler. Most installs are "drag this .app file to your Applications folder", meanwhile on Windows you download an installer that downloads another installer that does who-knows-what to your system and leaves ambiguously-named files and registry modifications all over the place.
- I worked at a company with a big, complex Postgres database behind the back end, probably half of all the incidents we had were related to the query planner suddenly deciding to change its approach.
Eventually you get a vague idea of how to coax Postgres to make the plans you want it to, but the fact that you can't at least lock it in and the plan might change with any number of factors at any time... That's just bad design.
- All I notice is that my time going from calendar to Teams call is ~30 seconds due to slow site loading and extra clicks. Calendar to Meet call is two clicks and loads instantly with sane defaults for camera/microphone settings. It's significantly better than teams or zoom in those regards.
- I found that with the bandwidth and storage that my company was using on the cloud, we could get ROI in under 2 months by building a server and running it in house. Now we've scaled up to a dozen servers but it's still just a handful of computers in a closet that saves us $50k/mo in cloud costs. It was dirt cheap to slap together and scale up incrementally.
- As a former Reddit mod I always found the self-promotion rules problematic. It effectively means you can promote your stuff all you want as long as you pretend you're someone else. It would be better to encourage people to stand behind their stuff. I tried not to remove self-promotion as long as it wasn't spammy (and there's a fine line there).
- I built a GPT-4 powered walking tour app during a hackathon. It uses the Wikipedia and Google Maps APIs to find interesting facts about things you're walking by and generates scripts to read aloud. It's fun and I learned a lot about places I go to all the time.
Unfortunately all the APIs are pretty costly so I can't post it for the group. Now that OpenAI's TTS and GPT-4 Turbo is out, I could probably revisit it. Still, not sure if people would be willing to pay enough for something like this.
- I started using Kubuntu and I absolutely loved it. Leagues ahead of Windows 11 in terms of user experience, customization, and speed.
However, I had to switch back. I just can't get by without Adobe apps and other specific productivity apps and video games that I couldn't get to run on Ubuntu. Maybe some day!
- According to [1], the new gpt-4-1106-preview model should be available to all, but the API is telling me "The model `gpt-4-1106-preview` does not exist or you do not have access to it."
Anyone able to call it from the API?
- A few notes on pricing:
- GPT-4 Turbo vision is much cheaper than I expected. A 768*768 px image costs $0.00765 to input. That's practical to replace more specialized computer vision models for many use-cases.
- ElevenLabs is $0.24 per 1K characters while OpenAI TTS HD is $0.03 per 1K characters. Elevenlabs still has voice copying but for many use-cases it's no longer competitive.
- It appears that there's no additional fee for the 128K context model, as opposed to previous models that charged extra for the longer context window. This is huge.
- I think it would be interesting to try a 14-day workweek, working 9 days on and 5 days off.
The advantage being that you can do a lot more on your weekends (international travel on a weekend is now possible), and during the week you'd have a lot of uninterrupted time to work. I imagine the productivity would be similar to a five-day week but it would feel like a lot more time off.
- It seems likely that the FTX creditors will be made whole or nearly whole. This is in large part due to FTX's investment in Anthropic that did a 5X[1].
1: https://coinscreed.com/ftx-creditor-claims-break-50c-as-buye...
