🤖 Made with AI
NVIDIA CEO JENSEN HUANG HAD SECOND ALLEGED AFFAIR WITH COMPANY EXECUTIVE
Nvidia ($NVDA ) CEO Jensen Huang reportedly had an extramarital relationship with Kimberly Powell, the company's VP of Healthcare, according to five former employees cited by The New York Times.
The two were reportedly seen being affectionate at Nvidia's March conference, raising potential conflict-of-interest concerns.
The allegations emerge as Nvidia's board investigates Huang's separate affair with a woman US intelligence identified as a suspected Chinese spy.
Nvidia declined to comment on the new report.
10
Before AI I could ship 1 medium or 3-5 small PRs/week. With Claude Code or Codex, I could get to 4-6/day without degradation in quality. I took Grok Bot, Cursor Cloud agents, pstack, my own skills, a eval suite generator, and retro tool, and I am pushing 20 high quality PRs/day and accelerating!
5
2
85
The "make me 20 variations" button made everyone in marketing feel like they were moving faster.
But it made no one better.
AI has yet to automate the important piece, knowing which variations will be the best.
Can you remember the last ad that made you say "WOW"?
1
29
The "make me 20 variations" button made everyone in marketing faster.
It made no one better.
We automated the part that was never the bottleneck. Knowing which one is right — and why — is exactly as hard as it always was.
What's the last piece of marketing that actually stopped you?
16
Everyone has the same AI tools now. Which is exactly why nothing anyone makes with them stands out.
The scarce thing was never generation. It's recognition — knowing which of the 20 is the one, and why.
So the real question isn't "can AI make the thing." It's: would you let it get inside your head?
Learn your taste, your references, the stuff you can't put in a brief — and finish your sentence better than you would have.
Most people flinch at that. Taste feels sacred, un-automatable.
I think the flinch is the whole opportunity.
The next great creative shops won't be measured by headcount. They'll be measured by taste per person — and that number is about to get very strange.
15
mustermania | AI Systems Architect retweeted
How to keep AI spend flat while token usage grows exponentially: Not with friction and spend alerts. With better defaults, routing, and caching.
Better Defaults (not Usage Caps) – Engineers can choose any model they want, but defaults matter. We’re experimenting with defaulting to open weight models like GLM 5.2 and Kimi 2.7 through our LLM gateway, while still encouraging engineers to choose the right model for the task. 91% of our employees were never hitting their usage caps, so instead of lowering caps and driving up alerts, we're moving to cheaper defaults. Note that code reviews use a diversity of models, so they can check each other's work.
Better Routing – In our custom harnesses, we preprocess prompts and route to the best model for the job, considering cache hits and model pricing. For instance, you may want a frontier model for planning, but not for execution where they can be overkill. Ultimately, humans shouldn't be choosing models - AI can automate this task.
Better Caching – Cache misses are the easiest way to drive your cost up. All of our requests are cache aware, so we’re reusing a warm cache wherever possible. For example, our cache hit rate went from 5% → 60% in LibreChat once properly implemented.
Keep Context Lean – Start fresh sessions when switching tasks. Scope file context narrowly. Disconnect unused tools. Don't just compact. The goal isn't fewer tokens used, it's fewer tokens wasted.
Better Visibility – Our engineers can use as many tokens as they want, from whatever model they want, but we’ve made usage visible – and the more you spend on AI, the more impact we expect.
The goal isn't to suppress usage. It's to build the infrastructure that makes exponential growth sustainable.
Putting this into practice has cut our AI spend nearly in half, while our token usage continues to grow.
The new model of entrepreneurship (2026+):
Build Audience → Discover Problems → Build Solutions → Go-to-Market
AI supercharges every step.
• Use AI to grow real audiences with smart content & engagement
• Mine comments & conversations to find real pains
• Prototype & iterate MVPs at lightning speed
• Launch with hyper-personalized messaging to people who already know you
Old way = guess and pray.
New way = audience-first + AI = dramatically lower risk and faster wins.
This is how solo builders and small teams win now.
Who’s operating like this? Drop your playbook 👇
#Entrepreneurship #AI #BuildInPublic
2
25
Obviously, most of these AI layoffs are just overcorrection for ZIRP hiring sprees. But… most companies, especially big tech likely will need to restructure and consolidate roles due to AI. Better to be a smart generalist who embraces AI systems than a specialized genius who fights AI.
NEW: $ORCL is cutting 21,000 jobs as AI adoption accelerates and credit risk flashes levels not seen since the GFC.
2
32
AI is going to drive major role convergence inside companies.
For decades, organizations optimized around specialization because execution was expensive — so we built clear handoffs between PMs, designers, engineers, QA, marketing, and sales.
AI is changing the cost equation. A strong engineer can now sketch, spec, test, build, and document. A strong seller can prospect, research, campaign, and create collateral. Not perfectly, but often well enough.
When execution gets cheaper than coordination, the advantage shifts to people who can work across boundaries.
Curious if others are seeing this too.
1
24
mustermania | AI Systems Architect retweeted
what if we kill the principal engineer and bind its soul to the llm?
I would like to make a proposal.
The hackers are currently sitting on Ethereum and Arbitrum.
They will use a number of "DeFi" things that have a number of single-points-of-failure.
Identify those single points of failure RIGHT NOW. Use them to intercept these funds and rug DPRK and get this fucking money back.
Literally EVERY. SINGLE. THING. they use has single points of failure. I guarantee it. 100% confidence. If you think otherwise you are in denial, stupid, need to get more creative, need to red team harder. I don't want to hear your fucking excuses.
Teams should do this right now. And then they should fully eliminate those single points so that they can't be used to steal money from anyone ever again.
Thank you for your attention to this matter.
The future of DeFi is not single nodes or single signer setups.
What we’re experiencing now is a lack of better security practices, which doesn’t matter 364 of 365 days of the year, until it does matter. A lot.
In this particular incident, you have basic security mismanagement that has been confirmed by sources close to the matter:
1.) It was an official LayerZero DVN that was attacked with very poor security practices.
2.) Applying these practices to a 1/1 DVN under centralized internal control which was exploited.
There was centralization risk on the amount of nodes (in this case, just one) and likely the way the DVNs were accessing the chain (through one or two RPCs).
Kelp relied solely on the LayerZero DVN. This is extremely irresponsible from a team with $1.5B in user funds under management. Unacceptable.
There are dozens more single DVNs out there that are still running with the same setup. For the 2/2 or 2/3 DVNs, its unknown how many of these are controlled by LZ themselves
Security researchers who have done diligence onchain close to this matter suggest that LayerZero runs a lot of these DVNs themselves. The official LZ DVN is setup as a 2/3 but all ran internally. So in the KelpDAO exploit, you had a single entity attack vector: LZ themselves.
There is a pattern of behavior that creates risk across all of the entire DeFi system, not just the LZ system. It’s all the tokenholders, issuers, and lending protocols that are now suffering from the design choices.
We've seen it time and time again.
These risks are completely unacceptable in bigger financial markets onchain and really sets us back in terms of adoption.
It also sets us back entirely as an industry built on the "don't trust, verify" mentality.
Misrepresentation of what your infrastructure is thus creating a massive web of risk around single node, single signer architectures across all of DeFi is a massive blow to what everyone is trying to achieve here.
We can do better. People need to be more informed about the risks they are taking. We need to be more rigorous about what we're calling decentralized.
Risk frameworks are coming to DeFi. Something has to change.
This is actually a spot on contrarian take.
If you start an app now, please don't use:
1. Supabase = you will regret it.
→ use Convex, Neon, Better-Auth instead
2. Clerk = this is the worst choice ever
→ use Better-Auth instead
3. Supabase Storage or AWS S3
→ use Cloudflare R2, this is the cheapest you can have
4. Namecheap: laggy, buggy, boring
→ use Porkbun or Cloudflare
5. MongoDB = just don't
27

















