Remarque’s LLM Lab retweeted
byok + grok + the latest claude and gemini models?
excalidraw is turing into a pretty serious ai playground 👀
and defaulting to gemini 3.8 flash for people who don't want to bring a key is a nice touh.
Remarque’s LLM Lab retweeted
480,126 deliveries. Consensus was 406,024. That's an 18.2% beat.
Eight years in finance - I've rarely seen a delivery report swing this hard versus expectations.
The context that makes this meaningful:
Q1 2026 was brutal. Tesla delivered just 358,023 vehicles in Q1 with a 50,000-unit inventory overhang after US demand cooled when EV tax credits expired. The market was pricing in a continued demand problem. Q2 answers that question definitively.
Model 3/Y carried almost all of it - 467,762 deliveries. Other models added 12,364. Total production was 451,758 meaning Tesla actually drew down the Q1 inventory build rather than adding to it. That's the cleaner signal inside the headline number.
Energy storage deployed 13.5 GWh - right on the ~13.8 GWh consensus - and that segment carries 39.5% gross margins versus 18.9% for automotive. Nobody's talking about that line but they should be.
What this doesn't answer: margins. Full financial results drop July 22. Deliveries this strong could still come with pricing concessions that compress gross margins. The delivery number is the volume story. The margin story is July 22.
TSLA was already up 8% before this report at $410. The demand recovery signal is now undeniable. Whether the margin structure supports the valuation is the next question.
BREAKING: $TSLA delivered 480,126 vehicles in Q2 versus expectations of 406K.
Remarque’s LLM Lab retweeted
Perplexity started as a small business tool for ourselves. We had 4 people and no revenue with AI at our fingertips.
The pivot to Computer is actually a full circle. Founders are using it to grow companies that matter to the economy and their communities.
It’s rewarding to see it now powering small businesses and startups in big ways.
Perplexity is still a startup. We just 5X’ed revenue from $100M to $500M with only 34% growth in team size. 2x revenue growth in 2026 with same small team. And we’re just warming up.
Everyone here works at a small business, and everything we build is for people who build.
Remarque’s LLM Lab retweeted
This is powerful.
Perplexity Computer + Plaid means your AI can finally work with real financial data, not just assumptions.
Remarque’s LLM Lab retweeted
Fortune: Perplexity CEO Aravind Srinivas just argued that AI-driven layoffs could push millions of people out of disliked jobs and into small, AI-powered businesses.
AI lowers the cost of starting a company because software can now handle work that once required teams of analysts, support staff, marketers, and junior engineers.
That changes the math of entrepreneurship because a founder can use LLMs, automation tools, and AI agents to write copy, answer customers, study competitors, build prototypes, and run back-office work with far fewer employees.
His view treats job loss as a temporary shock that opens a path toward one-person or very small firms that produce real revenue without the payroll that older companies needed.
But the problem with this scenario is that losing a stable paycheck is painful for most, and many workers cannot instantly become founders, and economists still disagree on whether AI is replacing labor at large scale or merely giving companies a new excuse for cuts.
---
fortune .com/2026/03/24/perplexity-ceo-ai-layoffs-not-bad-people-hate-jobs-entrepreneurship/
Remarque’s LLM Lab retweeted
That's it 😮
Microsoft is releasing Copilot Cowork to perform any task across your apps and files.
And they've collaborated with Anthropic to create it!
But you're not limited to only one model... which is a huge advantage compared to other tools.
Remarque’s LLM Lab retweeted
I built this onboarding before Claude Code or Codex existed. I made it open source on GitHub, so take it if you want.
Remarque’s LLM Lab retweeted
Today’s idea came from a real habit:
I order the same few items from the same restaurant… and somehow I still juggle Swiggy and Zomato to order..😅
So the question was: can I order in seconds? & can I do it without sending users through multiple screens?
Not sure it covers every real-world edge case, but it’s a solid experiment.
I studied Dynamic Island motion from Pinterest + Apple inspiration, mapped it in Figma, prepped the assets… and this is the result.
Remarque’s LLM Lab retweeted
if you want to journal every day in 2026, one year is a pretty good place to start 🌷



























