HN user

jeffnappi

566 karma
Posts4
Comments214
View on HN

These data points are based on 2023 inflated pricing as the entry point. Not much to see here.

In 2023 the car market was wildly overpriced due to the low interest rates and covid money combined with supply chain problems.

For Tesla owners in the U.S., their 2023 Model Ys are worth 42% less than what they paid two years ago

GPT-5 12 months ago

We have no idea what AGI might look like, for example entirely possible that if/when that threshold is reached it will be power/compute constrained in such a way that it's impact is softened. My expectation is that open models will eventually meet or exceed the capability of proprietary models and to a degree that has already happened.

It's the systems around the models where the proprietary value lies.

Linear is the best project management for software I've ever used, highly recommend. They've added many many amazing features this year... Incredible team over there that are just a joy to work with.

The problem is that folks who are hard of hearing want to be part of the conversation and make eye contact. The transcripts in the audio recorder app on Pixel phones is quite good, though combining computer vision and audio processing is necessary for loud and complex social environments.

This is great! Hearing aids are stupidly expensive and I hope we see more solutions leveraging mass production to improve accessibility of hearing aid tech

Another thing I'd love to see is affordable AR glasses with auto captioning for noisy environments with many speakers. My mother struggles with hearing loss, and it's apparent to me that audio-only solutions are not sufficient for many situations and types of hearing loss.

There's a small company already working on auto-captioning, but the price point is currently $5k - https://www.xanderglasses.com/xanderglasses

We were Urchin users prior to the acquisition and were also Google AdWords customers...

I remember being shocked at the idea of putting Google's JavaScript in our secure e-commerce site. We wanted to pay as little as possible for traffic and were mining the long tail of keywords. We certainly did not want AdWords to know what our economics were.

Ultimately AdWords evolved into a very efficient system that groups similar keywords and takes as much margin as the advertisers can handle. The golden days of a level playing field in ppc advertising were over and Google won the game :)

Vehicle to load type solutions will make the grid impact of EVs negligible. In fact over time it will help stabilize the grid and add storage for peak solar/wind generation. Imagine having most vehicles plugged in at an office building during the day for example, in a power outage the office building would be able to run off of the vehicles. Nissan in fact has such a setup[1]

1. https://www.nissan-global.com/EN/STORIES/RELEASES/nissan-wor...

Many things were platform specific at that time. You didn't have major game titles or business applications widely available across platforms. Folks who used PCs/Apple/etc at work/school were more likely to buy similar for the family at home etc.

In my family I get the impression we chose our home computers based on merit/value. That meant starting with the commodore 128 in the mid 80s and led to my brother buying an Amiga 1000 [with his hard earned teenage burger king min wage $] in 1987.

In the 90s the advent of Windows 3.1 running on cheap PC clones left Commodore in the dust. The value for the money shifted to PC, even if it was inferior at first.

It was really sad that Amiga did not continue to innovate - the hardware was astonishing which can be seen by looking at the demo scene output and games front he time relative to what was possible on other platforms.

Building open models is a very strong approach to cornering the market on top tier AI researchers. And as other commenters have mentioned, the raw models are not the product - the vast majority of the value is in how they are integrated into useful products.

This may be true, but as someone who contributed to a Terraform provider - HashiCorp did a vast amount of legwork to build impressive tooling and process to make contributing easy and get contributions merged quickly. No easy feat if you ask me. When I made contributions it was a matter of days before they were merged and released.