ahzhou

Fast as in time-to-market

VC vs bootstrap is usually based on company TAM. There are certainly high growth bootstrapped businesses.

Not common in Silicon Valley, but much more common in the rest of the country. There’s an archetype for bootstrapped tech businesses: - highly vertical specific - couple hundred million TAM - founder started the business in their 30s and is now in their 40s

It’s a tensor stored in GPU memory to improve inference throughput. Check out the PagedAttention (which introduces vLLM) paper for how most systems implement it nowadays.

They slightly restructure their MoE [1], but I think the main difference is that other big models (e.g Llama 504B) are dense and have higher FLOP requirements. MoE should represent a ~5x improvement. FP8 should be about a ~2x improvement.

We don’t know how much of a speed improvement GRPO represents. They didn’t say how many GPU hours went into to RLing DeepSeek-r1 and we don’t have a o1 numbers to compare.

There’s definitely lots of misinformation spreading though. The $5.5m number refers to Deepseek-v3, not Deepseek-r1. I don't want to take away from HighFlyer's accomplishment, though. I think a lot of these innovations were forced to work around H800 networking limitations, and it's impressive what they've done.

[1] https://arxiv.org/abs/2401.06066

You can easily do a fermi estimate based on the information given. They are comparing GPU hours.

See: https://planetbanatt.net/articles/v3fermi.html

I might be missing something, but DeepSeek’s recipe is right there in plain sight. Most of the cost efficiency of DeepSeek v3 seem to be attributable to MoE and FP8 training. DeepSeek R1s improvements are from GRPO-based RL.

Interesting to note - we have no idea how much R1 cost to train. To speculate - maybe DeepSeek’s release made an upcoming Llama release moot in comparison.

LLMs are inherently bad at this due to tokenization, scaling, and lack of training on the task. Anthropic’s computer use feature has a specialized model for pixel-counting: > Training Claude to count pixels accurately was critical. Without this skill, the model finds it difficult to give mouse commands. [1] For a VLM trained on identifying bounding boxes, check out PaliGemma [2]

You may also be able to get the computer use API to draw bounding boxes if the costs make sense.

That said, I think the correct solution is likely to use a non-VLM to draw bounding boxes. Depends on the dataset and problem.

1. https://www.anthropic.com/news/developing-computer-use 2. https://huggingface.co/blog/paligemma

Author: @fandzomga Username: fsndz

Why try to funnel us to your paywalled article?

Conditionally yes. There are many libraries that cannot be tree shaken for various reasons. Libraries typically need to stick to a subset of full JS to ensure that the code can be statically analyzed.

GraphQL is very powerful when combined with Relay. It’s useless extra bloat if you just use it like REST.

The difference between the two technologies is that LangChain was developed and funded before anyone know what to do with LLMs and GraphQL was internal tooling using to solve a real problem at Meta.

In a lot of ways, LangChain is a poor abstraction because the layer it’s abstracting was (and still is) in it’s infancy.

To be fair, LLM-based chatbots are much better about this because you don't need to discover the magic incantation to talk to a human. It's a trade-off because that same property introduces the possibility of hallucination.

It depends on the business, but the kind of metrics you are talking about are measured and taken seriously. People have absolutely gotten fired for CS quality KPI drops.

While it may not happen for you, “too lazy to look it up” is the vast majority of CS requests.

My understanding from talking to a couple of CS execs is that these have been a slam dunk in terms of ROI because CS agents don’t need to handle type C requests. I expect we’ll only see more as time goes on.

> AppStore would be dead on arrival

Certainly not. PMF was already established via the jailbreaking scene and Installer.app / Cydia. Millions of people went through the annoying processing of jailbreaking their phone to get apps.

If you’re saying that economics is a foundational driver of progress, then yes - almost by definition.

Banks and investors provide liquidity to the system, which is just one of many things the market demands.

Yes, this is a fundamental weakness with LLMs. Unfortunately this is likely unsolvable because the search space is exponential. Techniques like beam search help, but can only introduce a constant scaling factor.

That said, LLM reach their current performance despite this limitation.

They fall under a few buckets: Driver:

- node-postgres

- node-mysql2

Query Builder / Other thin clients: - knex - kysely - slonik ORM: - TypeORM - MikroORM - Objection.js - DrizzleORM - Prisma (actually runs a separate binary)

Its a two sided marketplace and companies only care about the conversion they get from different channels. If demand dries up, it will be reflected in more attractive pricing - I don’t think it’s likely that the entire market pulls out.

FWIW - it seems like the campaigns are working. You seem to be familiar with the brands and someone below chimed in on how one particular brand is great. Multiply that by the viewership - that’s definitely a win.

Some quick (unverified) research tells me that YouTuber marketing pays somewhere in the range of 30-70 CPM. You can pretty easily calculate that against google AdWords with reasonable conversion assumptions to decide if it’s worth it.

Css modules would be great, except there’s bad tooling in VSCode. Autocomplete through Typescript is the killer feature of Panda / Vanilla extract, not that you can style.

I was thinking about this too, but the wife of an actor and someone two years out of her masters were not the caliber people that should have been on the board of an $80B company.

I would expect people with backgrounds like Sheryl Sandberg or Dr. Lisa Sue to sit in the position. The two replaced women would have looked like diversity hires had they not been affiliated with an AI doomer organization.

I hope there’s diversity of representation as they fill out the rest of the board and there’s certainly women who have the credentials, but it’s important that they don’t appear grossly unqualified when they sit next to the other board members.

Firing a CEO is an extraordinary measure only taken in dramatic situations. Yes, the board can do it, but the legal and reputation risks are such that at the very least you let the CEO resign and pass the torch down.

You’re right - the board construction is remarkably bad. Seems like it was a classic case of centralizing power by stuffing the board with relative nobodies.

It’s incredibly common in every East Asian coastal cuisine.

If you live near the coast, you can easily forage your own too.

Mexico has been a victim of corruption since before it was a nation. Its political system under the Spanish was designed for efficient resource extraction and labor exploitation. Its home-grown system in the 1930's was created for the consolidation of political power and money to the winners of it's revolution.

Combine this with the fact that the drug market in the US is ~$150B, it's hardly surprising that the cartel pops up to service the demand while joining hands with the corrupt central government. Neither group cares about the people.

To begin to tackle it as you suggest, I think Mexico needs a full blown revolution which somehow manages to resist the billions of dollars the cartels are sitting on. It's not like Mexican's aren't trying either - it's just that if they try too hard, they get publicly executed.

Note that that herculean effort only addresses the supply problem. The demand will still be there until the US controls it's drug epidemic, which I don't think anyone knows how to do.

As I mentioned above, I'm not assigning any blame. I think both countries are in some kind of extremely shitty local maxima where the problems are so difficult to solve that a minority of people are able to profit and perpetuate the status quo.

I agree it's complex.

I'm not blaming the US. Like most problems in the world, it's a result of circumstance, not because there's a bad guy. It's not like there's a magical solution to market forces applying to drugs.

My point is more that the geopolitical reasons for the cartel's existence means that Mexico has much less agency in solving it than they would like.

I'm not even sure how Mexico manages to dig itself out. The only way to root out the cartels seems to just hope that the US figures out how to manage its drug market.

It's a pity - Mexico is so rich in culture but is just absolutely crushed by the US's demand for drugs.

I’ve never understood ESG funds.

1) It’s hard to reliably turn ESG goals into fair (non-gameable) portfolio metrics to incentivize funds

2) LPs still want funds to make returns, so even if they define good metrics, they still want most of the incentive to be based on returns.

3) ESG companies don’t post better returns than those from other asset classes.

It makes more sense for LPs put their money into non-ESG funds and just set aside some amount to achieve ESG goals through a charity with established metrics. They’ll get a tax break to boot.

Smart bulbs are a PITA anyways. If you just need on/off/dim (most people) a Lutron Caseta based system is way more reliable and you don't need to worry about unintuitive light switches.

> Why not spend the money building better power plants?

This is equivalent to the “build it and customers will come” fallacy of startups. You still have to convince humans that it’s all worth it.

The US could have had the nuclear industry of France (~70% of its electricity generation) if we had only gotten everyone to agree in the 80s. At the end of the day, you need money and marketing to convince people.