I build things | Dm for work/collab
- Tweets 13,374
- Following 687
- Followers 16,890
- Likes 26,540
Pinned Tweet
I spend most of my time exploring new AI products, developer tools, and interesting tech.
If you're building something worth people's attention whether you're an indie developer, startup, or global brand I鈥檇 love to explore working together.
Here's how I approach collaborations:
- I choose quality partnerships over quantity.
- I only work with products I would genuinely use and recommend.
- I am always interested in early access to AI models, beta programs, launches, APIs, SDKs, and developer tools.
- As a developer, I can build demos, experiments, or mini projects around your product to help people understand what makes it useful.
- Happy to share honest feedback on both software and hardware.
- If your product is physical, I'm happy to test it.
- Always open to merch as well. 馃槃
I'm especially interested in AI, developer infrastructure, open source, and tools that help people build faster.
馃摡 pankajams1234@gmail.com
馃挰 DMs are open.
10
8
66
53,729
OpenAI Gonna Reveal a Long-Term Agent at DevDay
- This agent would focus on long-running tasks, similar to Grok Bot or Manus, where it can keep working for hours, days, or even weeks.
- I think it could be built on Astra, which is already quite good at long-horizon tasks.
- The codename is "Aeon", and it designed to stay persistent.
- Its gonna have a cloud environment, similar to Cursor. It would set everything up in the cloud and work on the task there, continuing for hours, days, or weeks until the task is completed.
- OpenAI already has the infrastructure for long-running sessions, hosted sandboxes, and multi-agent workflows, so this would make sense as the next step.
- I have also heard it could have multiple agents talking to each other and working together on the same task. If OpenAI actually ships that, things are gonna get crazy.
18
12
84
6,765
Opus 5 is really amazing at motion design.
I gave it one prompt: make a 30s film where the camera keeps zooming into letters, and every letter reveals a new word.
18 words, all hidden inside each other.
The typography, animation, camera moves, music, beat sync, it made everything itself.
One prompt. Around 50 minutes.
1
841
Pankaj Kumar retweeted
Opus 5 is really amazing at motion design.
I gave it one prompt: make a 30s film where the camera keeps zooming into letters, and every letter reveals a new word.
18 words, all hidden inside each other.
The typography, animation, camera moves, music, beat sync, it made everything itself.
One prompt. Around 50 minutes.
8
5
34
3,430
Opus 5 is really amazing at motion design.
I gave it one prompt: make a 30s film where the camera keeps zooming into letters, and every letter reveals a new word.
18 words, all hidden inside each other.
The typography, animation, camera moves, music, beat sync, it made everything itself.
One prompt. Around 50 minutes.
8
5
34
3,430
Opus 5 is really amazing at motion design.
I gave it one prompt: make a 30s film where the camera keeps zooming into letters, and every letter reveals a new word.
18 words, all hidden inside each other.
The typography, animation, camera moves, music, beat sync, it made everything itself.
One prompt. Around 50 minutes.
1
244
Pankaj Kumar retweeted
Opus 5.5 is insanely good at motion design.
I gave it one prompt: "create a motion design video of a poster breaking out of its own frame."
The animation, the camera moves, the sound design, even the music. It composed the whole soundtrack itself and synced every hit to the beat.
Sound on 馃攰
5
3
33
3,147
Sonnet 5.5 Also Isn't Too Far
- Sonnet 5.5 is already in testing, and some partners have received access for 24 hours.
- If testing and feedback continue to go well, I think it could launch next week or in the first week of October.
- My expected date is around September 30 or October 1.
- It will have the same 1M-token context and 128K max output.
- Thinking is adaptive by default, with Low/Medium/High options. No xhigh.
- Pricing: $2 input / $10 output per 1M tokens, with $0.20/M cache reads, same as Sonnet 5.
- Opus 5.5 is a very good model, so I have high hopes for Sonnet 5.5. I mainly want better token efficiency + Opus 5.5-level output quality, without unnecessarily generating huge responses.
Anthropic is currently stealth testing Claude Sonnet 5.5 (`claude-sonnet-5-5`).
Check comments for more information meow
8
9
156
16,608
Pankaj Kumar retweeted
I鈥檝e set aside $20M because I want more startups to grow with us.
Higgsfield crossed a $1B revenue run-rate today. We鈥檙e celebrating this with 100% cashback on Higgsfield API spend with Genjutsu and Cinema Studio available.
When you're a small startup, early support can change how quickly you build.
Special thanks to @gregisenberg for showing how to get started with GPT-6 Astra + Higgsfield API, step by step.
I hope this helps you ship faster and grow into one of our biggest customers.
We're announcing 100% cashback on every model on the Higgsfield API platform.
Seedance 2.5, Kling 3.0, MiniMax H3, Wan 3.0, and more.
Spend on the API and get your cashback instantly, up to $100,000 per business.
$20,000,000 cashback pool. First come, first served.
You helped us reach a $1B run rate. We鈥檙e celebrating by putting $20M back into what you build next.
Unused cashback expires on September 30.
55
196
541
80,641
Pankaj Kumar retweeted
Tested a stealth model called Space Bunny with one prompt: turn a resume into a retro game.
Fed it a fake test profile, and it built a playable career map with 3 unlockable levels, player stats, and a final-boss skill tree.
Single HTML file. Free on OpenCode this week
6
2
18
4,056
Tested a stealth model called Space Bunny with one prompt: turn a resume into a retro game.
Fed it a fake test profile, and it built a playable career map with 3 unlockable levels, player stats, and a final-boss skill tree.
Single HTML file. Free on OpenCode this week
6
2
18
4,056
Sonnet 5.5 Also Isn't Too Far
- Sonnet 5.5 is already in testing, and some partners have received access for 24 hours.
- If testing and feedback continue to go well, I think it could launch next week or in the first week of October.
- My expected date is around September 30 or October 1.
- It will have the same 1M-token context and 128K max output.
- Thinking is adaptive by default, with Low/Medium/High options. No xhigh.
- Pricing: $2 input / $10 output per 1M tokens, with $0.20/M cache reads, same as Sonnet 5.
- Opus 5.5 is a very good model, so I have high hopes for Sonnet 5.5. I mainly want better token efficiency + Opus 5.5-level output quality, without unnecessarily generating huge responses.
3
392
Pankaj Kumar retweeted
Claude Opus 5.5 built a 300 metre tall flower that blooms in real time.
Day-to-night skies, dynamic rain, fireflies, and even a cutaway view.
Zero assets. Just code.
Cost: 10M tokens.
8
4
49
3,442
GPT-6 Sol & Luna Are Officially Out
- GPT-6 Sol: $2 input / $10 output. Luna: $0.10 / $0.50 per 1M tokens, 50% cheaper than GPT-5.6 pricing.
- AutomationBench: Sol 33.2% at $0.27/task vs Opus 5 26.9% at $3.00/task.
- DeepSWE: Sol 68.8% vs Fable 5 69.9%, at around 80% lower cost per task.
- OSWorld 2.0: Sol 60.5% vs Opus 5 60.3%, at ~80% lower cost.
- They handpicked the benchmark notably using Fable 5, not Fable 5.1.
- And yeah, Sol isn't as good overall as Opus 5.5, that's why they reduced pricing.
Overall, a good upgrade for the price.
2
2
14
3,280