arnon

8 ms·

61.
by arnon 11mo ago
build things """"people"""" want
62.
by arnon 11mo ago
So I can't have opinions? Also, this is entirely hand-written ;)
63.
by arnon 11mo ago
I guess that's true for some but not for all. I wouldn't say that's the most common scenario
64.
The trust collapse: Infinite AI content is awful(arnon.dk)
237 points by arnon 11mo ago|210 comments
65.
Computational Complexity of Air Travel Planning (2003) [pdf](ai.mit.edu)
81 points by arnon 11mo ago|12 comments
66.
$2.7B Agent Tax Crisis: First-ever study on how AI agents avoid tax(paid.ai)
1 points by arnon 11mo ago|0 comments
67.
Notes on where seat-based pricing is going(arnon.dk)
1 points by arnon 1y ago|0 comments
68.
Notes on where seat-based pricing is going(paid.ai)
2 points by arnon 1y ago|0 comments
69.
by arnon 1y ago
If I am a company that builds agents, and I sell it to someone. Then, that someone loses money because this agent did something it wasn't supposed to: who's responsible? Me as the person who sold it? OpenAI who I use below? Anthro
70.
Legal Contracts Built for AI Agents(paid.ai)
72 points by arnon 1y ago|45 comments
71.
Trump says Michael Dell is part of the team buying TikTok with Ellison, Murdochs(theregister.com)
6 points by arnon 1y ago|0 comments
72.
by arnon 1y ago
That's what I'm advocating for
73.
Why your outdoorsy friend suddenly has a gummy bear power bank(theverge.com)
258 points by arnon 1y ago|319 comments
74.
Forward Deployed Engineers(svpg.com)
1 points by arnon 1y ago|0 comments
75.
Few of Waymo's most serious crashes were Waymo's fault(understandingai.org)
5 points by arnon 1y ago|2 comments
76.
by arnon 1y ago
And ARR assumes it's recurring too - it's a metric that (people incorrectly think) shows how good your business is at retaining revenue year-over-year. If a big proportion of contracts in Y1 get terminated at the end of the 12m pe
77.
by arnon 1y ago
Usage-based will be fine, but if you sell usage-based with a discount that assumes the customer will stay with you for, say, 3 years - this law means they can leave after 1 and not stay for the full length. Then, you've given a discoun
78.
by arnon 1y ago
Agreed, but it does open up some cans of worms where you may accidentally lose customers when they forget to renew (even if they meant to)
79.
by arnon 1y ago
ARR assumes customers will stay with you - that's why it's a bad metric.
80.
by arnon 1y ago
It also applies to "data processing services": "Article 25, Contractual terms concerning switching … The rights of the customer and the obligations of the provider of data processing services in relation to switching between
81.
by arnon 1y ago
The issue is you can't use the "I'll give you a deeper discount for a 3 year contract" line anymore. It'll potentially create a situation where customers in the US pay a lot less because they still have that incenti
82.
The EU Just Killed ARR(paid.ai)
75 points by arnon 1y ago|94 comments
83.
Are you selling agents the way customers want to buy?(paid.ai)
1 points by arnon 1y ago|0 comments
84.
Chaos Is the Operating System(paid.ai)
2 points by arnon 1y ago|0 comments
85.
MIT's 95% AI failure rate is wrong(arnon.dk)
3 points by arnon 1y ago|0 comments
86.
Vibe Revenue – a mirage of AI success(paid.ai)
3 points by arnon 1y ago|0 comments
87.
The free tier death cult(arnon.dk)
2 points by arnon 1y ago|0 comments
88.
by arnon 1y ago
Agree - not every company is going to """vibe code""" their own ticketing solutions - it's just not an integral part of running their business. Having a more tailored, "AI-native" solution is, we
89.
by arnon 1y ago
He's setting a dangerous trap for others - just like they did when they said they were firing all human support and then rehired them later. Yes, Klarna can rebuild everything because they have hundreds of engineers and effectively inf
90.
AI agents lose money every second they run(paid.ai)
3 points by arnon 1y ago|0 comments

More ›