Yash.

A product marketer tinkering with AI as Ferrari loses another F1 race.

[Yash. @Aweracker_Yash]

Product Marketing lesson from OpenAI DevDay: when your flagship model fails safety review the week of your biggest event, you don't cancel. You rename agents "dots", ship a bubbly avatar, and let the mascot carry the keynote.

iPhone Air announcement when everyone came for the Vision Pro.

Same day, the $200 plan quietly went from 20x to 10x. Ship the mascot, slip the changelog.

[Yash. @Aweracker_Yash]

@hailleymari Instinct is doing this for me. It's literally saving my ass so many times. It evens texts me to remind me I had committed this in that meeting 😂

[Yash. @Aweracker_Yash]

So this is true. With the IPOs beckoning and data going public.

Anthropic's leaked data says they made $4.6 billion in earnings on the back of $42 billion in losses. This will only continue if AI companies keep bleeding money on acquisition and retention.

Frontier models are amazing for marketing use cases, but 95% of today's AI use could be done with Sonnet, GPT-5.6 Sol, or DS 4.1 and a mix of other closed or open-weight models.

Once these companies IPO, everything becomes public, and such losses might finally rupture the bubble, which OpenAI and Anthropic don't want at all. The momentum they have with each new model's introduction is bringing big gains. If their financials put the brakes on innovation, that's where things will start going wrong.

Overall, the new strategy is built to move to profitability and doesn't actually say the days of subsidies are gone. But remember today's 6-Luna is as good as yesterday's Opus 5 or GPT 5.5-Sol.

Every frontier model from Dec 2025 is now a mid-level model; that is the pace. So your usage will change, but overall, work quality improves even with smaller models.

[Tibo @thsottiaux]

Hi,

Tomorrow we are re-opening the Pro $200 subscriptions to new subscribers, but together with it we are also changing how we calculate the usage for it. In effect, if you do the math, it will net out at half the dollar in API spend compared to the old Pro $200 plan.

Now that it's said, let me explain why this is happening and why you will still get more work done than if you were on the Pro $200 subscription one month ago.

(a) We didn't want to compromise in other ways and are committing to not reintroducing the 5h limit, so that you can fully use the weekly usage when you want.
(b) On the subscription, we guarantee that over time you always get more work done and with an increasing level of quality. This means that you will continue to get more value per dollar spent as a result of models getting more efficient and us passing down the improvements in the form of API price reductions.
(c) We don't want to put an incentive on ourselves to artificially inflate the API list prices to make it look like you are getting a lot (and workaround it through discounts, etc). Instead we want to continue to both rapidly reduce prices and increase capabilities of models on the API. This week we introduced GPT-6 Sol and GPT-6 Luna at 50% of their previous price. Over time, we see prices go low enough that it makes sense for most to buy usage as needed without there being a significant gap between what you get in a subscription and what you get in the API for a dollar spent.
(d) Tomorrow, we are adding more things to the subscription that won't draw on the usage, I won't reveal what that is yet.

I wanted to be transparent before all the big announcements tomorrow. Lots of new exciting things are coming to the subscriptions that will make it super compelling, but I wanted to make sure to share this change ahead of time so you can all understand it before we shower you with good news.

Codexingly,
Tibo

[Yash. @Aweracker_Yash]

Every CX leader on a panel says the same thing: customers want to talk to a human.

No they don't. They want their refund.

Think about your own life. You have never once contacted a brand because you craved connection. You contacted them because something broke and you wanted it unbroken. The "human touch" is what people ask for after the machine has failed them twice. It's not a preference. It's a symptom.

The real hierarchy:

Solved in 30 seconds by an AI > solved in 30 minutes by a human > an empathy statement from anyone.

Nobody has ever received "I completely understand your frustration" at hour 26 and felt understood.

Here's why this take makes people mad: half the industry's business model is the humans. Of course the keynote says customers love them. Meanwhile every eCom brand with a working AI agent watches its human tickets drop and its CSAT go up. The data has been in for a while. It just ruins the panel.

AI in CX was never about replacing empathy. It's about replacing the wait. The brands that get this stop performing care and start delivering it.

Your customers are not lonely. Your NPS is.

[Yash. @Aweracker_Yash]

Every AI demo is a marketing team showing you the one prompt that worked.

[Yash. @Aweracker_Yash]

Everyone wishes they had a Dad like Lance Stroll. Good for nothing themselves but Dad's got a racing team and I have a lifetime contract.

[Yash. @Aweracker_Yash]

I gave Claude a hand-drawn diagram of my flat's floor plan and asked it to simulate a house design.

After a few rounds of back-and-forth, I have a 3D model of my actual 2-bed flat that I can furnish with some furniture I can bring in:

What it does:

- Furniture, gadgets, and utilities are listed, and I can add more if needed for simulating how the home might look.

- Double-tap to pick things up, drag them around, put lamps on tables, swap finishes.

Day and night, lamps on or off, and a dimmer on every lamp.

- A ray-traced mode, so I can see how light falls in each room before buying a single lamp.

Works on my phone too.

@claudeai @ClaudeDevs

And it runs in the browser: https://t.co/f4mpO0hjuq

Same reminder as always: AI is the worst it will ever be.

[Yash. @Aweracker_Yash]

@Jefe_OK @LandoNorris @LandoNorris

[Yash. @Aweracker_Yash]

@LandoNorris Talks about mental health and promotes healthy mental health. Yet he says Colapinto should be race-banned and "I don't know why some folks are in F1". Dude you have also made these mistakes. Remember Austin?

How insensitive when you say Colapinto has just 2 years in F1. And he apologized straight away. He knew he fucked up.

[Yash. @Aweracker_Yash]

@paraschopra I have been using Claude for around a year for video gen, but Opus 5.5 and Astra are both so good at it.

This is what I did yesterday eve - https://t.co/pzxe5NyC6D

[Yash. @Aweracker_Yash]

Opus 5.5 is egregiously good at making videos. Have been making videos with AI for almost a year, but this is incredible.

We did a product video this week, funnily, without any deep knowledge of how video could be made manually.

It's a 75-second film about how you train Richpanel's AI: you give it your website and your SOPs, then you coach it in plain English, the way you'd onboard a new teammate.

It's not perfect, and I'm sure a real motion designer would spot plenty. But going from "I don't make videos" to something I'm fine sharing, in about a day, still feels a bit strange.

That's also the point of the video: you shouldn't need to be an expert to get something good out of an AI. You just need to know what good looks like and say it clearly.

@claudeai #opus #gemini

[Yash. @Aweracker_Yash]

Opus 5.5 is egregiously good at making videos. Have been making videos with AI for almost a year, but this is incredible.

We did a product video this week, funnily, without any deep knowledge of how video could be made manually.

It's a 75-second film about how you train Richpanel's AI: you give it your website and your SOPs, then you coach it in plain English, the way you'd onboard a new teammate.

It's not perfect, and I'm sure a real motion designer would spot plenty. But going from "I don't make videos" to something I'm fine sharing, in about a day, still feels a bit strange.

That's also the point of the video: you shouldn't need to be an expert to get something good out of an AI. You just need to know what good looks like and say it clearly.

@claudeai #opus #gemini

[Yash. @Aweracker_Yash]

Wut...

[Xiaomi MiMo @XiaomiMiMo]

Introducing Xiaomi MiMo-V2.6 — Pro & Flash.

Frontier intelligence, all the modalities, built in public.

🔹 Two omnimodal models, advancing through scaled reinforcement learning
🔹 Pro performs on par with Claude Opus 5 and GPT-5.6 Sol across most agent benchmarks
🔹 Pro scores 46 on the Artificial Analysis Intelligence Index — the highest among open-source models
🔹 Stronger coding, computer use, 3D reasoning and creative capabilities
🔹 Open model weights, technical report, RL environments and training code

Blog：https://t.co/JGdwC9Ymvj

[Yash. @Aweracker_Yash]

Why is this still a thing?

[Yash. @Aweracker_Yash]

How much AI do you use?

Me: Yes!

[Yash. @Aweracker_Yash]

No one can figure what the next steps would be. It's just a cluster F everywhere. Yikes.

[voxium @v0xium]

I am done with this shit. It is over. The state of engineering right now is horrible. It has been half a month since I started a new role at a big company. Nobody knows anything here. The specs, code, tests, PRDs, tickets, resolution of those tickets, reports, etc., everything is made by Claude Code. Nobody on my team likes this. They are being forced to ship as much as they can. I have heard multiple times from higher management that pushing code is not a bottleneck, so why are we slow? People are working 12 to 13 hours a day just to press enter. Nobody is reading anything. Humans in corporate are doing nothing on their own. Everyone, literally everyone, from an L1 to an L7 engineer here is doing the same thing. Talk to Claude. There is no sense of victory. Nobody is resolving bugs. In reality, nobody is thinking anymore. Everything is done by LLMs. It is so soul-sucking. I would not mind it, to be honest, if we were at least given the time to check out the code and see what is going where. But no, the goal is to just ship. No matter what happens.

[Yash. @Aweracker_Yash]

Link to repo - https://t.co/nAuGr5IjuM

[Yash. @Aweracker_Yash]

Thanks to TipTour and Jev Browser, I was able to vibe code a small tool just to learn what Jev can be useful for.

[Yash. @Aweracker_Yash]

Tweeting from JEV

[Yash. @Aweracker_Yash]

@thsottiaux Usage Limit.