adampk

Do you mean that there is no way to value OpenAI for a purchase so a sale is impossible?

At the least OpenAI is worth what the market is willing to pay for it.

So strange, does the author think companies never try to understand their customers?

I don't get why a normal corporate internal resource system its being framed as ridiculous. Does the writer not know about "personas"? Weird tone.

I worked through this for a tax company. They had a huge pile of artifacts from tax questions worked up for clients. What we did is we "reverse engineered" the process of the questions that would lead to that tax memo and the research steps to find the sources and conclusions. It worked well and we were able to replicate the process which the SME's created these memos.

For a given tax question, could you come up with the same memo quoting the same sources and same conclusion?

You would think this would be obvious to everyone. Clearly Apple is prepping for a digital overlay on the real world. Also less UI interaction, more voice/AI interaction.

How easy is the set up, does this need to be deeply integrated in each step of the life-cycle?

It does mention that, it calls that out specifically.

As you grow, it’s tempting to fix every issue using the ‘cowboy’ method. It’s fast. It’s efficient. It leads to good results. But the number of things that need a cowboy fix grow exponentially, and cowboy fixes only ever fix that one thing, while system fixes fix future issues too. As you adapt from cowboy to drone, it’s easy to skew too much to one side or the other. No matter how good your systems are, sometimes stuff just needs to get done pronto. But sometimes you need to take a step back and trust that the system you built will do its job, and trying to jump in to speed things up will only make everything worse.

Vertex, Inc. - Commerce Solutions | AI Product Engineer | REMOTE | Full-Time https://vertexinc.wd1.myworkdayjobs.com/VertexInc/job/Remote...

We are hiring an AI Product Engineer to join the strategy team to use the latest and greatest in AI to push Vertex forward into an AI-first company.

If you want to test drive being a technical founder, have experience building the full stack of an AI product from 0 to 1, and want to make a dramatic impact on in a public company, please apply.

You introduce a point I have not seen discussed before which is that these type of content distribution platforms go through a process to find their global minima.

Twitter at the beginning you didn't know what it was going to be or what worked. Same with facebook and instagram. As time goes on these sites small features bring out their emergent properties of what 'works' there.

And once it has been 'figured out', it is not as fun. You know what you can expect there and people go there but it is no longer a dynamic feeling. Like watching the NBA today, it has been 'figured out'.

I think that may be what is the factor in the longevity of these platforms, once it is 'figured out', if what it is, appeals to enough of a large base.

Tik tok may have gone further because it never really was 'figured out' in that larger way. The algorithm really could give you wildly different content and different 'trends' would show up so it never reached that static boring point.

For these 'on the decline' sites you can almost predict exactly what you will see there and exactly what the discussions are. It is not longer an exciting TV show.

Vertex, Inc. - Commerce Solutions | AI Product Engineer | REMOTE | Full-Time

https://vertexinc.wd1.myworkdayjobs.com/VertexInc/job/Remote...

We are hiring an AI Product Engineer to join the strategy team to use the latest and greatest in AI to push Vertex forward into an AI-first company.

If you want to test drive being a technical founder, have experience building the full stack of an AI product from 0 to 1, and want to make a dramatic impact on in a public company, please apply.

'you put more bits into the prompt, you get more bits out.'

I think your line there highlights the difference in what I mean by 'insight'. If I provided in a context window every manufacturing technique that exists, all of base experimental results on all chemical reactions, every known emergent property that is known, etc, I do not agree that it would then be able to produce novel insights.

This is not an ego issue where I do not want it be able to do insightful thinking because I am a 'profound power'. You can put in all the context needed where you have an insight, and it will not be able to generate it. I would very much like it to be able to do that. It would be very helpful.

Do you see how '“superintelligence” is nothing more than a(n extremely) clever arrangement of millions of gpt-3 prompts working together in harmony' is circular? extremely clever == superintelligence

I don't disagree that there is never a need for 'new data' to make progress. But there is plenty of novel engineering that can be done with 'new data'. Just needing insights and connections.

But realizing that you can use certain commodity devices or known processing techniques in different problem spaces does not require new data, just 'insight'.

Agreed! Don't get me wrong, the statistical distribution modeling for human language is still SUPER helpful. And for things like legal/tax/coding, which has a lot to do with applying language patterns, this is a very big deal. But the ability to find the true 'sub structure' of content that it is trained on is not something they can do. It is like there is some lower substrate that it is 'missing'. That is a lot to ask for, but once we get there it will be the 'magic' that is promised, rather than amazing, super helpful, parlor tricks.

Interesting, could you give me an example? LLMs definitely can "understand" what I am asking at times when a human couldn't. They have more data to 'find similarity' to what I might mean. But I do not think you are saying they answer questions a human couldn't?

The closest example I could think of is the (maybe true, maybe myth making) story of SpaceX using car wash valves instead of super expensive 'space grade' valves that did the same thing, and were orders of magnitude cheaper. Doesn't seem like embodied AI is necessary to figure this out.

Yes! The ChatGPT moment has warn off. And there hasn't been a step-change other than Claude Sonnet 3.5 + Cursor for dramatic impact (which is only for coding) since then.

I 100% agree with you that AI is fantastic and it is a big deal in general. But now that the world has gotten used to it being able to parrot back something it learned (including reasoning) in the training set, the next 'big deal' is actual insight.

But I see your point, I still think what we have currently is out of a sci-fi book, but I am also not that amazed by computers in our pockets anymore :)

This is the big idea in the paper, basically that CoT is limited for some complex problems because there is a class of problems where there is no 'textbook' way to find a solution. These are novel problems that need a unique methodology. "Essentially, to start generating the solution requires that we already know the full approach. The underlying generative process of the solution is not auto-regressive from left-to-right."

Mathematical meaning:

We can formalize this argument through the interpretation of reasoning as a latent variable process (Phan et al., 2023). In particular, classical CoT can be viewed as (equation) i.e., the probability of the final answer being produced by a marginalization over latent reasoning chains.

We claim that for complex problems, the true solution generating process should be viewed as (equation) i.e., the joint probability distribution of the solution (a, s1, . . . , s) is conditioned on the latent generative process. Notice that this argument is a meta-generalization of the prior CoT argument, hence why we will refer to the process q → z1 → . . . → z as Meta-CoT.

I think this is seminal. It is getting at heart of some issues. Ask o1-pro how you could make a 1550nm laser diode operating at 1ghz have low geometric loss without an expensive collimator using commodity materials or novel manufacturing approaches using first principle physics and the illusion is lost that o1-pro is a big deal. 'Novel' engineering is out of reach because there is no text book on how to do novel engineering and these class of problems is 'not auto-regressive from left-to-right'.

In your view, does any company that does business with government organizations count as sucking on the government teat?

I did not notice any implication of Americans being incapable of innovation. It is wrong and, frankly racist, to imply immigrants are used to suppress wages.

What is the "killer app" for something that would take this enormous level of effort?

If you work more hours and take more overtime, you will make more money you can use to build wealth.

Seems like a clear link.

> Retrieval Aware Fine-Tuning (RAFT), presents a novel recipe to prepare fine-tuning data to tailor the models for domain-specific open-book setting, equivalent to in-domain RAG.

I cannot see the insight on why this is a for a limited domain? The key problem that is being solved is the known problem where RAG returns an irrelevant chunk. It seems like the "benefit" is training a model to ignore irrelevant chunks.

I am guessing because it costs money to train on multi-domains so they limited their research on one-domain at a time but not sure if there is a "bigger reason" why this isn't an approach to a fine-tuned "make answers from only relevant chunks" model? The paper seems to imply this is only works for specific-domains but I can't see why.

What is the best place to use hosted versions of these models?

We want to interact with it same as the OpenAI API.

Where did they openly lament?

Yes it would be very hard to make a bad technology choice when you thought through the whole system

It clearly explained what the design constraints were and why the technology choices met them

"That data came to Meta through its Pixel code, which the tax firms installed on their websites to gather information on how to improve their own marketing campaigns. In exchange, Meta was able to access the data to write targeted algorithms for its own users."

Seems that they provided the data to get Meta to target customers for them.

When I hear animation I think of 3D character rigging in a 3D engine with animations. This seems to be text-to-video?

I agree with you. A "everything is terrible" campaign. But who stands to gain?

Taleb makes the point that this is not as simple as it seems. It is not the same people staying rich.

https://medium.com/incerto/inequality-and-skin-in-the-game-d...