Posting stuff about tech and products, maybe books and random too. Tech product manager and 2x founder. Still code, mostly with agents these days.
San Francisco Bay Area, CA
Joined November 2007
- Tweets12.9K
- Following2.9K
- Followers1.4K
- Likes20.8K
Pinned Tweet
My take on this: if you use Obsidian (or Notion or Notes or whatever-you-use) for primarily keeping track of to-dos, your 'second brain' is going to end up being your 'anxious brain'. Rather, using it for keeping a journal (to revisit the old self sometime in future), or hard to narrow down nuggets of info, or keeping track of some research - are use-cases that would make you want to preserve it for long.
buff.ly/hPHyoV1
Manish retweeted
You should be changing your mind about AI at least every 12 months. zou-group.github.io/llm-clai…
Noting that Mistral publishing this directly on alphaXiv is a key milestone for alphaXiv. I see the trend flipping on where the discussions originate - this is an improvement.
"Mistral Large 4"
paper published exclusively on alphaXiv
Their latest technical report shares that aggressively reducing KV-cache memory can hurt an LLM's agentic capabilities, even when its language modeling performance looks unchanged.
Mistral initially reduced KV-cache memory by 16x, but they found that this hurts long-context agentic tasks, where models need to retrieve information from much earlier context.
To fix this, Mistral restored full-context attention to more layers, then adopted DeepSeek Sparse Attention to select the 2,048 most relevant tokens across the context instead of attending to everything.
Their resulting 1T-parameter MoE model (w 52B active) scores 41% on Terminal-Bench 4.0 and 19.7% on FrontierCode 2.0, and its NVFP4-quantized checkpoint fits on four GB300 GPUs.
alphaxiv.org/abs/2611.mistra…
Agree with this. Even more relevant in current (and incoming) state where you can't bank on the market moat remaining as-is later for you to capitalize. Build workable flywheel now, and improve / adapt it as you go through the trajectory.
If you are building a startup, try to be profitable when you are small.
We turned ebidta positive on a ~$2.25 Mn ARR.
Focus on product led growth.
Built clear edge with product.
~50% user acquisition was organic, it still is.
Had a small team. Small office.
100% work from office, it helps.
Extremely low marketing spends.
Focus on unit economics, from day 1.
When you scale a startup and become large with zero profits, it becomes extremely difficult to turn things around. Not impossible, but is super challenging.
Scale now, profitable later works very rarely.
This is superb!
Introducing Reality-Aware Rendering in @Simulon, a completely new way to create with the real world.
Place neural assets with your phone and act out the scene. When you stop recording, Simulon renders the take, reading everything in it (what you do, what you say and the place around you), and your assets respond to all of it.
The final is regenerated from your recording: a new performance of the scene, not a copy of the footage.
Available now on iPhone. Coming to Android soon.
Manish retweeted
Underrated life advice: Be careful what you normalize around your kids. How you speak to your spouse. How you treat your body. How often you complain. How you talk about others. What feels ordinary in your house becomes how they'll act as adults outside of it.
Manish retweeted
Still no customers, but I’ve narrowed it down to the border radius.
Fixing that now.
Manish retweeted
PHP changes are live. Start patching. php.net/archive/2026.php#202…
Huge thanks to coordinated work by @ThePHPF!
PHP sites behind Cloudflare already have an initial set of WAF rules deployed for the known exploit vectors in these vulnerabilities. The rules are available to all customers, including free.
We are seeing early attacks emerging and are updating mitigations as new vectors show up.
This is a STOPGAP only. Patch as soon as @ThePHPF releases the update at 14:00 UTC today.
Good launch and clear intro. This is going to become critical for consumer-side apps as well (built into Phone OS, like Caller ID functionality).
today we're opening @DetectIfAI to every developer.
real-time voice deepfake detection, 24 ms per check.
#1 in latency on Podonos' independent benchmark, ahead of companies that raised hundreds of millions.
97.3% of cloned voices caught.
API + MCP. link in the comments.
Manish retweeted
Tech people in the Bay Area want to save time but the rest of America just wants to save money
Manish retweeted
Vibe coding is not an excuse to stop thinking.
It’s possible to agentic coding as an intentional tool instead of a slot machine.
Don’t use AI to surrender your cognition, use it expand what’s possible.
Manish retweeted
What if,
for any scientific claim,
you can search 100M+ papers to find the strongest evidence for and against it
⚡️in 1 second?
That’s now possible with Evidence API! 🚀
paperclip.gxl.ai/evidence-ap…
For example, does drinking coffee reduce diabetes risk?
