More enterprises are running a portfolio of models. Easy tasks go to smaller, cheaper models, while frontier models take the hard tail. Hosting that portfolio is a GPU utilization problem. The models don't fit on a single node together, and 9-34 min cold starts make switching between them impractical. Excited to show our work with NVIDIA's Dynamo team. We show how to serve a portfolio of frontier coding models on a single 8x B200 node, switching models in ~1 minute with in-flight sessions intact. cedana.com/blog/coding-model…

As the Pareto frontier of model cost vs. capability creeps up, every sophisticated token consumer is going to be multi-model. @cedanacorp dramatically improves the performance of sharing compute infra for a model portfolio. cedana.com/blog/coding-model…

People who drank more coffee had less body fat, less visceral fat, and more lean mass than less-frequent or non-drinkers. The biggest benefit was seen at 5 or more cups per day. And for men, every extra cup per day was associated with a ~10 ng/dL higher total testosterone. Higher coffee intake was also associated with better glucose and insulin regulation in men. This was an observational study, but mechanistic and experimental evidence actually supports coffee's benefits on metabolic and hormonal health. I'm willing to believe there's a real link here rather than draw a conclusion such as "coffee drinkers live healthier lives in general."

The secret behind Japan’s perfect school lunches

The age of the 40-year-old founder is back. Bryant Chou spent 12 years as CTO of Webflow, which now powers something like 1.5% of the entire internet. He's back in the current YC batch with Ploy, an AI marketing platform, and he describes himself as "a bit of a boomer, double the age of the YC founders." But over 13% of his batch is already using his product, within months of launch. There is a side of the argument which destroyed one of the main edges young founders have, which was being faster and cheaper at building.... speed is everyones game. If what's left is knowing what to build this is more likely to come from spending 15 years watching an industry up close, collecting the thousand small frustrations that tell you where the real problem is. Bryant can build an anti-slop website tool because he spent over a decade learning exactly why websites are slop. So I'm updating. I don't think it's young vs. old. I think AI rewards whoever has the most domain knowledge to point it at, and only sometimes is this younger founders who are thinking outside of the box...

Secondary markets are rife with fraud and bad actors, and it pains me to see these bottom feeders profiting off Anduril's growth while fleecing retail investors through unreasonable or opaque fee structures. In this week's episode of nonsense, @teamignitevc (a fund we've never taken a meeting with or had any contact with whatsoever) founded by @brianrbell (who we've never met) is soliciting investors via a public Google Doc to invest in an SPV that will in turn invest in another SPV that will in turn potentially enter into a Forward Contract with a (supposedly, though unnamed) early Anduril employee: docs.google.com/document/u/0… Few problems here. First off, so-called "Forward Contracts" are notoriously hard to settle in private companies, and counterparty risk is extremely real (e.g. what about the many complicating corner cases like acquisitions where shares don't "trade" or marriages/divorces/deaths where ownership of underlying shares is complicated)... just generally a risky structure to close that I don't think most folks actually understand. Second, this "Deal Memo" includes basically no details about Anduril's performance, no revenue figures whatsoever, no product specifics... almost as if it's soliciting investors to invest on hype and momentum and not fundamentals. Generally I'd advise folks to be skeptical of any "deal memo" lacking basic details. Third, Forward Contracts are explicitly disallowed by Anduril's stock plan and bylaws, as we detail here: anduril.com/investor-relatio… ... which means that Anduril will never consent to Team Ignite's SPV actually taking possession of these shares while we are privately held. Zero chance. And finally, the memo spends most of it's time talking about the structure and fees... which are INSANE. A double-layered SPV with all legal and administrative costs passed through, in addition to an 8% upfront fee, 3% annual fee for two years, 20% carried interest, and... the craziest part... an implied price per share that is completely insane (in this case, the implied pps is ~115% higher than the most recent preferred raise from like 9 months ago). Flattered, I suppose, but also puts these investors in an almost absurd position by paying more than double the pps of our most recent transaction. 😬 As stated at the top, I don't know Brian or Team Ignite at all, maybe they're kind wholesome people and this is all a big misunderstanding... but if I were an investor looking at this "opportunity," I'd run for the hills.

I’m not an investor in @KianSadeghi5, but my hot take is that he’s a great founder trying to figure out how to deliver a new service to a market. It doesn’t feel like he has cut corners and it sounds like the biggest push back is on his marketing which was supposed to provoke people — so in that they got more than they bargained for. He must be on to something, because people seem pissed about it.

Top things I look for in an investor (as a YC founder) 1/ Writes checks quickly I’ve never met a great investor who did not write checks quickly. The better the VC the faster they work, on average. A big reason for this is they respect founders time. They realize that the most sought after founders want to get back to business and want partners who respect that. I’ve also found some of the more nightmarish VCs to force long, convoluted and drawn out processes. So definitely a pattern and speed is a good first check. I always ask: “So if things go well here, what happens next? can you walk me through your typical investment process” 2/ Not “hands on” I’ve met a lot of investors who brag to me about “value add” and “hands on” assistance. That is not what I’m looking for. I don’t want an overly confident MBA student diving into my business thinking they know what they’re doing. Can investors help? Sure. But outside of capital they won’t be make or break for my company. Personally, I want to minimize the number of people in my kitchen that don’t know how to cook. You can fund the restaurant but you’re not “helping” the chef - this can turn from helpful to harmful quickly. 3/ Previous founder Exception to all of the above is previous founders. Across the board they are phenomenal to work with. They understand the life, the pain, the ups/downs. They even know to get out of your kitchen when it’s too crowded. Outside of fast checks (which previous founders tend to do anyways), this is my second favorite criteria. Previous founders are more often than not a pleasure to work with. 4/ Not a sheep Many VCs simply tag along in the same rounds with their friends. This follow-the-leader behavior leads to a lot of signaling required from the founder. I prefer to work with firms who can make a decision independently, using their own logic. These folks tend to write first checks (first is most important), follow on if you need capital in the future, and move quickly. 5/ Good reviews Yup. I background check if I can. Calling up portfolio company founders to get a read on what it’s like working with the firm. How did they act when things were going well? What happened when things went poorly? Did they ever do anything shady or not-founder-friendly? There’s been a few instances where a friendly investor was flagged by other founders for doing unfriendly things. — If an investor meets these criteria and willing to invest on reasonable terms, I wouldn’t over think it. Take the check, get back to business. This is my personal list, not a golden rule. you may have a different criteria you look for. What did I miss?

I appreciate you are being honester about the numerator (some ppl here claiming it’s only 65k) But you aren’t quite being honest about denominator: H1Bs occupy good jobs Per grok they take 20-30% of jr software jobs at big tech. Which is much easier to see on your grid