Fractional CRO.
Zero Reach
Joined January 2011
- Tweets7.3K
- Following213
- Followers1.1K
- Likes8.7K
Market always pushes back against pay-per-value models. It sounds 'too expensive', so people just don't pay, even if incentives are aligned.
Hence, 9-5 salaries, hourly rates in freelancing etc.
Everything pushed toward something, that applies downward pressure on the price of output, even if two hours and outcomes are not equal.
That's the most common framing on the market.
This, obviously, works differently in ad space, where you actually need to compete on getting the right outcome (a lead or a sale) with other businesses, not necessarily individuals.
So if AI providers manage frame it that way, they might also succeed in running this type of payment models.
The top models could do independent research and go to companies with capacity of developing findings into positive financial outcomes, and let them bid against each other to have access to the chain-of-thought data.
All intellectual property rights will be forfeited by the AI model provider, if someone pays for it.
Provides a real value-incentive feedback loop, where to look next for better payoff too.
Dario Amodei: some tokens Claude produces are worth a few cents, others could be worth tens of millions of dollars.
His examples:
1. Telling someone to restart their Mac: the model has said it 10 million times, worth a dollar or less
2. Telling a pharma company to move an aromatic ring to the other end of a molecule: potentially tens of millions
3. Expect pay-for-results pricing, or labor-like pricing by the hour
4. It's a new industry, so many models will be tried
Today pricing is simpler: per the S-1, usage-based pricing was 80% of H1 2026 revenue and subscriptions 20%.
Full breakdown here:
theaiopportunities.com/p/ant…
So AI space is speedrunning Telco fiber buildout, where people benefited from it, but the companies investing money into it 'due to demand' struggled.
And this will have a much faster feedback loop.
GPUs in data centers have shelf life of a few years.
Fiber can last decades.
Pretty much.
As of the past few weeks, debt markets have almost entirely shut off access to money for any company that doesn’t have substantial revenue independent of AI data centers.
SpaceX closed $40B of debt for GPUs, but we’re investment grade and have Starlink & launch revenue.
That means the 200+ neoclouds are going to see guppy-level death and consolidation. Some will make it through, but not many.
If there's a day when everything will go down because of a massive cyber attack on essential infrastructure, it will be on GTA VI release day.
IMPOSSIBLE CHALLENGE:
Surviving Black Friday without getting your account cleaned or identity stolen this year.
The workflow now is:
> hypothesis
> let AI pull in years of order and customer data
> pull any third party data needed and event specific timelines that could be relevant for denoising
> let multiple frontier AIs do analysis with and without being aware of the hypothesis (2-pass is better)
> test or implement, if supported by most
A lot of others findings in the process.
Once Opus 5.5-class+ layoffs start rolling in, if you're into cheap nice-to-have SaaS, you'll be fighting steep churn.
People won't have money to spend on that stuff.
Pivot to B2B, because it's strategic.
There's a lot of automation to build out (which would kill your SaaS customers' wallet).
And, because they will also be losing customers for the same reason, they'll need your help to both cut expenses and get business going.
There's no way to lose.
Would you rather have:
> 300 $10/mo SaaS customers
> 1 $3K/mo B2B client/customer
Now you might want to jump to the conclusion that B2B setup is no different than a 9-5.
But I haven't mentioned:
> how many hours either takes a month
> how demanding it is mentally
> how churn and customer retention looks like
> marketing required to get there
There is a point at which you might realize that consolidating into a smaller, better paying client segment comes with a lot of 'quality of life' benefits.
Especially, if you're a solo founder.
It doesn't need to be the 300 vs 1 extreme, but somewhere in between there's a right balance of having a good time.
Well, at least if you enjoy slower days and peace of mind.
A lot of the reasoning behind a $50+ pricing floor is segmentation.
People who will pay $50 would also pay $10. But not many who will pay $10 would also pay $50. And those “cheap” customers are often the most demanding.
You CAN get to $50k MRR on $10 subs (I did with FeedbackPanda), but is SO much more work on the customer service side. And often, low-budget customers are extremely demanding and unreasonable.
So, you segment them out. You set an entry price tier at $50 ( I did that initially with Podscan, later moved to $100 for even more filtering) and get rid of interest (and workload) from that segment.
Also, very importantly, professional customers (who buy for their job) are much more likely to have a 3-4 figure budget for tooling. You can expand into existing teams much better than from one individual customer to another unrelated one.
Reducing prices tends to hurt revenue and your sanity. Increasing prices does the opposite in most cases.
(And for the extremely price-conscious customers, you can always lure them in with significant discounts.)
Alright, I'm turning the internet off until it's rewritten in Rust.
BREAKING: OpenAI says it found a critical flaw in Brotli, a widely used compression library built into software like Node.js, Bun and Envoy.
The bug could let an attacker take over a system just by sending it malicious compressed data, with no login needed. That works both ways: a hacker could hijack an unpatched server, or a malicious server could hijack an unpatched user’s device.
Didn't you say we won't get a Opus 5.5 class open weight model within 6-months, @thekitze?
I think it's only been about a month now.
Even if this isn't it, I'm sure we can get there in 5.
Introducing Kimi K3.1 Pro: our most capable 3T-class open-weight model.
🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal.
🔹 Full-project coding: games in Unreal, Unity & Godot.
🔹 Excels at parametric 3D modeling in Blender & CAD.
🔹 More reliable at finance work: Accurate modeling with hallucination down by 65%.
🔹 Matching the strongest proprietary models across coding, agentic and knowledge benchmarks.
Kimi K3.1 Pro and K3.1 Pro Highspeed (2–3× faster) are now live on Kimi.ai, Kimi Work, Kimi Code, and the Kimi API.
Open weights by Nov 4, 2026.
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.ai/blog/kimi-k3-1-pro
So all the 'good' engineers are either avoiding AI, using DeepSeek to destroy the Internet, or trying to escape the permanent underclass with vibecoded slop.
You know what, he's probably right.
We have entered the drone warfare age in cyber.
Reject AI and let people with unlimited DeepSeek compute to take everything from you.
It's okay. You'll die pure.
PHP, Node.js, 4
All dropping like flies. It will be a fun decade ahead.
🚨 BREAKING: Node.js just dropped its OWN security bombshell.
4 HIGH-SEVERITY vulnerabilities patched TODAY, including one that could let hackers remotely execute malicious code on vulnerable servers.
→ 10 SECURITY FLAWS were addressed in today's Node.js release, FOUR rated HIGH severity.
→ REMOTE CODE EXECUTION: CVE-2026-111497 is a serious flaw in Node.js's Brotli decoder that could allow attackers to execute malicious code remotely.
→ TLS SECURITY BYPASS: Two separate vulnerabilities could trick affected applications into accepting certificates for the WRONG server under certain conditions.
→ Think about that. Your application could establish an encrypted connection while failing to properly verify WHO it's talking to.
→ THREE MAJOR NODE.JS VERSIONS AFFECTED: Node.js 22, 24 and 26 all required emergency security updates.
→ RUNNING NODE.JS 18 OR 20? Those versions are END-OF-LIFE and aren't receiving these official security patches.
→ PATCH IMMEDIATELY: Update to Node.js 22.23.4, 24.22.1 or 26.12.1, depending on your release line. If you're using Docker, rebuild and redeploy your containers.
PHP AND Node.js both facing major security disclosures on the SAME DAY.
If you run a website or application, TODAY is the day to check your servers.
And remember: the public release of security patches can give attackers a roadmap to systems that haven't updated.
Don't be the last one to patch.
I oneshot a prototype for a product in <2 hours a week ago.
A person belonging to ICP was very impressed.
Even started asking technical questions, if it "does X, Y, Z" etc, but it was just a quick prototype.
Now it's a real project to be deployed in a few weeks probably.
However, a week later, when the enthusiasm would have usually faded:
I'm actually puzzled... why isn't this a thing already used by many?
It's a common frustration in the industry, and the leading market player could have solved it too.
Yet, there's still an opportunity.
Or, perhaps, it doesn't work?
The ICP might like it, but maybe the customers of theirs do not?
I personally use it and find it very convenient, but I'm biased. Or am I?
Time will tell, but this thing started making money right as the prototype was shared.
Luck appreciates real problems being solved.
