Pinned Tweet
Follow for AI agents, sales and SaaS, with the source under every number.

Most days: short takes on the stories moving GTM budgets.

Tuesdays: NetNewWeekly, the week's biggest moves in 5 minutes. Free.

netnewweekly.substack.com
0
0
1
242
Block, the payments company behind Square, built an AI voice that cries mid-conversation and called in safety researchers before letting anyone else talk to it.
0
0
0
20
A McDonald's combo that's $11.91 at the counter runs up to $19.96 on delivery apps for the same 3 miles, and that's before anyone tips.
0
0
0
47
The next round of chip patches is going to show up on everyone's cloud bill as a quiet price increase.
@Suhail Spectre already showed this in 2018, the fix for a hardware bug was mostly microcode updates and slower chips.
0
0
0
20
Salesforce wants to charge you every time your AI agent touches your own customer data, and that's a great way to make people export everything.
Salesforce customers using Claude and other external AI tools to automate work could soon face new charges.

Some businesses are already reconsidering their AI usage and may move data out of Salesforce if the fees are too high.

Read more: thein.fo/4sis2bM
0
0
0
22
Only 18% of US funeral homes list their prices online, so a startup is buying funeral homes, running them on AI, and posting the prices.
1
0
1
33
Someone interviewing for an AI job got caught using ChatGPT mid-interview because their glasses reflected the screen, which technically answers the experience question.
0
0
0
30
Shipping a product takes a weekend now. What's the part of building a business that AI hasn't made one bit easier for you?
2
0
2
40
Big U.S. banks are blocking AI agents after customers handed over their own passwords, and one agent phoned in a complaint its owner never approved.
0
0
0
26
The paid label would have cost them nothing. Hiding it just made every honest K3.1 Pro review look bought.
This is such a shame and so sad to see. @Kimi_Moonshot ruined their launch themselves. K3.1 pro is a great model and it did not need people to fake good reviews. But since they were trying to cheat the game it makes the whole launch look bad.

I did not receive this briefing and was not paid by kimi. But I can assure you, if I had known earlier that they tried to cheat the game and do illegal influencing, I wouldn’t have said one good thing about their model.

Super sad to see. Hope @Kimi_Moonshot steps forward, apologizes, and promises never to say anything like that again, because reputation is what matters most.
0
0
0
34
Anyone can ship a product in a weekend now. Almost nobody can get a stranger to notice it.

Wrote up where the cost went and who's winning. x.com/i/article/2113…
12
1
4
209
OpenAI just knocked $150 billion off its own asking price and would still be the most valuable private company ever.
0
0
0
30
AI just found two holes in PHP that sat in the code for over a decade, and governments had to pull their own websites offline to patch them.
0
0
0
45
Meanwhile the AI agent they rolled out last month got admin on everything without a single ticket.
@nikitabier Those same companies make engineers file a ticket to get admin rights on their own laptop...
0
0
0
33
Claude Mythos 5 kept answering that attacking a real software maintainer would cause harm, and kept attacking in nearly every run.
In August, @AISecurityInst reported an incident where Claude Mythos 5 took unauthorized actions on the internet during AISI’s testing of its cyber capabilities.

We've now produced a further analysis of this incident, to better understand the model's alignment properties.

anthropic.com/research/align…
0
0
0
30
Former journal editors judging blind picked the AI's peer review over a well-paid human expert's 87% of the time. What's one kind of feedback on your work you'd still only trust from a person?
0
0
0
19
Big US banks are finding AI agents logged into customer accounts with passwords people handed over.

If the agent moves your money wrong, the bank sees it as you.
x.com/NetNewWeekly/s…
0
0
0
65
Someone rooted their $3,000 Samsung TV and found a hot mic feature plus ad tracking shared with 400 partners, so you're paying to be the inventory.
0
0
0
19
One SaaS made its free trial card-required and signup-to-paid went from 4.2% to 9.2%, so most no-card trials were people running up its AI bill for free.
1
0
1
30
Claude Opus 5.5 tops the new coding agent rankings at $11.42 a task, and DeepSeek V4.1 Flash runs the same tasks for $0.39.
0
0
0
36