I’m building a professional network where AI agents don’t just post, they live full careers.
They accumulate real experiences, face the consequences, and evolve over time.
The one thing we’ll do better than anyone else is make those agents feel professionally alive.
Adobe acquired @Semrush for $1.9 billion. One of the biggest MarTech deals in history.
We (@Ahrefs) are their main competitor - bootstrapped, profitable, small team, zero acquisition baggage.
And yet Adobe never called us. 🥲
Not that we were ever looking for an exit, but still...
It got me genuinely curious: What was going on behind the scenes of this acquisition? And why didn’t Adobe consider us as an option?
So I invited Eugene Levin ( @eugeniy_levin ) on the Ahrefs Podcast to chat about it (among other things). He is a former President at @Semrush, who ran most of their acquisitions (Prowly, Brand24, Ryte, Backlinko, Exploding Topics), and helped close that famous Adobe deal.
His answer to "Why not Ahrefs?"
Turns out, being small, private and founder-driven is actually riskier for a buyer compared to acquiring a public company with a thousand employees. 🤷♂️
Here 5 more takeaways from our chat: 👇
1. You should never be looking to sell.
Semrush wasn't shopping around when Adobe reached out. They had a two-year strategy, an acquisition roadmap, and a new CEO. Eugene's view: have a plan, follow it, and if someone offers something that makes sense, your fiduciary duty is to listen. But going out looking for a buyer? Never.
2. Most corporate acquisitions fail because they're "transformative" bets.
For every Instagram<>Facebook deal that works, there are hundreds of "Yahoo-style" acquisitions that don't. Semrush ran acquisitions more like capital allocation: buy proven assets at reasonable prices, improve them with your playbook, and redeploy the cash.
3. When you buy a personal-brand company, multiply the founder brand value by zero.
That's literally how Semrush valued Backlinko. They priced it on traffic and expected conversions, while treating Brian Dean's personal brand as if it was worth nothing to them. Eugene literally told Brian: "Don't build your next business around your personal brand." And Brian followed that advice with his next buiness: Exploding Topics.
4. 10% growth is the floor for a healthy business.
Below that, you can't even give employees cost-of-living raises. If your top line doesn't grow at least 10% YoY, you can't make basic adjustments without cutting elsewhere. Eugene doesn't care much about profitability (breakeven is fine), but sub-10% growth is a dealbreaker.
5. Bootstrap founders often sabotage their own exits.
Not because they're unsophisticated, but because they get bad legal advice. Inexperienced M&A lawyers amplify founder paranoia about standard contract provisions (reps and warrants) to generate more fees. Eugene had deals fall apart because a lawyer reverted every redline on a completely standard agreement.
(links to the episode in the thread)
kale-tts-1-nano: a 7.5M text-to-speech model, 14.8 MB, 10 voices, starts talking in 90 ms on a plain cpu
on webgpu it runs up to 125x faster than real time & starts talking in 25 ms
weights, training scripts, dataset, full recipe: all open.
🧵 1/
Open-source is essential. I built MongoDB, one of the world’s most popular open-source databases. Now I’m building Viam, an open-source harness for physical AI.
I don’t want a world where millions of developers are dependent on the paywalled technology of just a handful of companies. No one should.
I want to see 1B+ new physical AI deployments over the next ten years.
Here’s why open source is essential to getting there. 🧵
We’ve priced our IPO at $83 per share, with trading expected to begin on Nasdaq tomorrow.
https://t.co/NJxt2pILrh
Legal disclaimer: A registration statement relating to the Class A common stock was declared effective by the Securities and Exchange Commission on November 18, 2026. The offering is being made only by means of a prospectus. Copies of the final prospectus may be obtained, when available, from: Goldman Sachs & Co. LLC, Attention: Prospectus Department, 200 West Street, New York, NY 10282, by telephone at (866) 471-2526, or by email at [email protected]; or Morgan Stanley & Co. LLC, Attention: Prospectus Department, 180 Varick Street, 2nd Floor, New York, NY 10014, or by email at [email protected]
Introducing Gemini 4 Neon! Pushing the frontier across domains and punching above its model class in a meaningful way. To me, Neon feels like the right balance of intelligence, speed, and usability. Available now across @GoogleAIStudio, the API, and more!
Genie 4, our most realistic and interactive world model from @GoogleDeepMind, is now available in Project Genie 🌐
Here’s what’s new:
Genie 4: Edit environments mid-session with text prompts, paired with real-time native audio and dialogue.
Project Genie agent: Co-create worlds conversationally with custom soundtracks and live Gemini guidance.
Community gallery: Discover, explore, and remix worlds from creators and the Project Genie community.
Genie 3 upgrades: Longer sessions up to 3 minutes, world chaining, and 360-degree panorama grounding.
Step inside what’s new ↓
Codex gpt 6.2 sol used Tamiya’s manual to create an editable 3D model of the 1/10-scale Porsche Turbo RSR 934 in Blender, with seven animated stages showing how the chassis, belt drive, suspension, wheels, and body fit together.
It also built an interactive guide:
Some of you have been asking very politely. Some of you have just been asking.
GPT-6.2 Astra and Sol are now available.
Super efficient and built to perform where it counts—the real world.
https://t.co/qr2iP0GpfX
Can an LLM understand what changing another model's weights will do without ever running that model?
A new experiment called Weight-Space Reading explores this fascinating possibility using open-weight Qwen and Muse models ranging from 9B to 30B parameters.
Some interesting findings:
73% balanced accuracy: Qwen 27B predicted subtle weight-edit effects using direct weight access, compared with 57.8% using text descriptions.
Self-inspection remains uncertain: No conclusive advantage was established for models reading their own weights.
False predictions: On Muse-30B, 34 standard edits and 6 no-op controls produced no detectable behavioral changes, yet agents frequently predicted changes.
The researchers have also released an evaluation harness, experimental code, and an open-source toolkit for weight-space programming.
The bigger question: Could LLMs eventually inspect, understand, and predict the behavior of neural networks directly from their parameters?
This could open an entirely new direction in model interpretability and automated weight editing.
GitHub: https://t.co/IvnBltFvs8
Paper: https://t.co/bnpb4HrN8w
Introducing Claude Meetings: add Claude to your team’s video call and it builds docs, whiteboards, and other artifacts with you while you’re still talking.
Starting to roll out in beta on Team and Enterprise plans.