fra
fra on Hacker News
Posted by
fra on Tuesday October 20, 2026 @05:44PM
from the blog.baldassari.me dept.
For all intents and purposes Character.ai was acquired in August 2024, and this is just the mess leftover from their run around antitrust regulation.
The company filing for chapter 11 today is shell that was left after Google "hi-quired" (the opposite of acqui-hired) the founding team. The shareholders got paid ($2.7B from Google), the founding team moved on, and what's left is the corporate structure and whoever Google did not want.
It sucks for the people left over (but then again, it also sucks to be part of the team that's laid off in a merger). It's also a terrible practice that I hope the Justice department and the CFTC will put an end to soon.
BERT need to be fine tuned for your use case, Jev generalizes. It’s a pretty big difference!
Most valuable comment here. You can fine tune BERT to perform really well on classification tasks. What's exciting about Jev is that it generalizes. Laya does not!
Interesting result, but the writing is very poor.
I agree that from the perspective of a business owner this would seem very normal. "I need legal services for X, and I happen to know and trust a very good outside counsel for this" is very normal. Additionally, law firms have super robust conflict of interest policies so you can expect this would have been reviewed in details.
However, I think this story brings up two important topics:
1. It is important that high level government officials avoid not just actual conflicts of interest but also the APPEARANCE of a conflict of interest.
2. The Trump administration has taken a much more liberal approach to conflicts of interest than most admins, and this has undermined trust in government. The fact that this story got people so riled up is a great illustration of it. This is not good for the country!
The demand signals never allocated the resources, a small number of people always did. Those people react to the demand signals (and they still do).
IMO a more important question is: what happens when the market power of a segment of society grows much greater than the rest? In a "K-shaped" economy, the top 20% have so much spending power (and even more future spending power) than their demand signals can overwhelm the rest. Then IMO the market won't respond to the demand signals of the bottom 20% and capitalism stops working for all. Add to that the possibility of AI automating more and more work and you have a system that only works for folks at the tops.
This isn't a resource allocation problem! it's a purchasing power distribution problem.
Less progressive than in the USA, believe it or not…
I would take the other side of that bet. In five years, there won’t be more outages than today.
I think most business will decline to do any further business with you after you threaten to sue them. There are even patent licenses that automatically get voided in the event you sue your counterparty for any reason.
the number of companies using local AI for development is too small to draw any conclusions from.
I have a simple list: money, medication, identification, underwear, bathing suit. Everything else you can figure out.
Deals of this type are usually cash free and debt free. The cash went to the shareholders. Hence the implied valuation of $2.5B.
I’ve seen many people climb the SWE ladder or build a YC company (having done both myself), and trust me the founder route is not the most efficient. You only believe this because you didn’t see the 2 failed startups and 10 year journey grinding 80h weeks almost running out of cash a few times.
Jensen Huang himself says nobody in their right mind should start a company https://www.cnbc.com/amp/2024/05/11/jensen-huang-i-didnt-kno...
For two reasons:
1. They create externalities
2. They generate an economic flux that can be taxed
My experience could not be more different. I’ve made life long friends at work, especially when I was working for smaller firms. I don’t think those relationships are transactional.
If you follow his logic and believe that the ultra-wealthy pay too little tax (as e.g. Warren Buffett does), then a balanced approach is to set the tax rate to: "37% of income or 1.85% of wealth, whichever is higher".
This would close the gap between Buffett's tax rate and that of his secretary, but would not be the "highest taxes in the world" that PG decries.
Posted by
fra on Saturday May 02, 2026 @01:37PM
from the blog.baldassari.me dept.
This was my first thought: this seems like a web + 3d port of Fritzing. Perhaps they even reuse some of the code?
Gross margin doesn't tell you much about their level of investment. Gross margin is only revenue minus COGS (i.e. hosting, support, potentially infra teams). To understand further investments you'd have to know R&D or at least Opex broken out for Apple Services (which AFAIK they do not share).
SaaS typically expects 80% gross margin, so Apple is not out of line here.
In practice, I don’t think tax policy is the primary driver for where people chose to live. Even very wealthy people.
If I were very wealthy the first thing I would want to buy is proximity to the people I love, the food I like, the culture I’m excited about. For many people that means being close to home.
I can’t imagine how shallow my life would have to be for me to optimize it around taxes.
Extremely impressive that they were able to ship inside-out tracking, pancake lenses, and eye tracking + foveated rendering. Each of these is a serious engineering challenge. Very few organizations could pull this off.
This guy ^ runs a conference on corporate values. You can't make this stuff up...
Hacker News formula for startups: no offices, no offsites, no meetings, and no MBAs. If only idiot CEOs and rapacious VCs were listening!
In hindsight I think the correct word would be "regular software upgrades".
Every vendor you bring in comes with switching cost, and this is no different. But we do our best to make it easy for folks to migrate off of our platform should they decide.