sdesol

> The LLMs are much more eager to please and to write lots of code.

My process right now when working LLMs is to do the following:

- Create problem and solution statement

- Create requirements and user stories

- Create architecture

- Create skeleton code

- Map the skeleton code

- Create the full code

At every step, where I don't need the full code, the LLM will start coding and I need to stop it and add "Do not generate code. The focus is still design".

> you are not being lousy you still need to think about all the edge cases

This is honestly where I believe LLMs can really shine. I think we like to believe the problems we are solving are unique, but I strongly believe most of us are solving problems that have already been solved. What I've found is, if you provide the LLM with enough information, it will surface edge cases that you haven't thought of and implement logic in your code that you haven't thought of.

I'm currently using LLMs to build my file drag and drop component for my chat app. You can view the problem and solution statement at https://beta.gitsense.com/?chat=da8fcd73-6b99-43d6-8ec0-b1ce...

By chatting with the LLM, I created four user stories that I never thought of to improve user experience and security. I don't necessarily think it is about knowing the edge cases, but rather it is about knowing how to describe the problem and your solution. If you can do that, LLMs can really help you surface edge cases and help you write logic.

Obviously what I am working on, is really not novel, but I think a lot of the stuff we are doing isn't that novel, if we can properly break it down.

So for interviews that allow LLMs, I would honestly spend 5 minutes chatting with it to create a problem and solution statement. I'm sure if you can properly articulate the problem, the LLM can help you devise a solution and a plan.

> I haven’t tried o3, but one issue I struggle with in large context analysis tasks is the LLMs are never thorough.

o3 does look very promising with regards to large context analysis. I used the same raw data and ran the same prompt as Simon for GPT-4o, GPT-4o mini and DeepSeek R1 and compared their output. You can find the analysis below:

https://beta.gitsense.com/?chat=46493969-17b2-4806-a99c-5d93...

The o3-min model was quite thorough. With reasoning models, it looks like dealing with long context might have gotten a lot better.

Edit:

I was curious if I could get R1 to be more thorough and got the following interesting tidbits.

- Depth Variance: R1 analysis provides more technical infrastructure insights, while o3-mini focuses on developer experience

- Geopolitical Focus: Only R1 analysis addresses China-West tensions explicitly

- Philosophical Scope: R1 contains broader industry meta-commentary absent in o3-mini

- Contrarian Views: o3-mini dedicates specific section to minority opinions

- Temporal Aspects: R1 emphasizes future-looking questions, o3-mini focuses on current implementation

You can find the full analysis at

https://beta.gitsense.com/?chat=95741f4f-b11f-4f0b-8239-83c7...

> Are people really locked out of LLMs due to price?

I think so. When you start using a good model, you soon learn that it can make a difference. If you have domain knowledge, you can treat LLMs like a design partner. With DeepSeek, what would have cost 50 dollars to use Sonnet is now 5 dollars. Having to only spend 5 dollars a month can be a real game changer for many.

This isn't about the China now. This is about preventing people from providing/using DeepSeek. Basically, this is about ensuring DeepSeek only has a market in China or in other China friendly countries.

Yeah I took a look at some of them and they are okay. I may need to switch to them as DeepSeek doesn't look like they can keep up with demand.

> Deepseek the company is not related to the weights of the model.

I know the model is there, but running the 671B model would literally not make sense unless you can provide services at scale. The only reason that I, and others care is DeepSeek costs 15 times less than Sonnet. Stopping DeepSeek from offering their service at that price is the whole point of why OpenAI and Microsoft is now serious about all of this.

I believe the goal is to prevent any other provider from providing the models as an API service. OpenAI and Anthropic charge as much as they do because they bake in high salaries and R&D. With the weights freely available, the only cost is just hardware and support staff.

At the price point that DeepSeek is charging?

Well, it looks like the ball is in the politicians' court right now, and who knows where this will lead. I said it in another comment, but I hope others in China and around the world will start to use v3 to build better models and by the time any legislation comes into play, it would be too late.

> There is literally nothing that the government can do to stop anyone from hosting the model.

I admire your optimism but the simple fact is, you have a person in power that can be easily bribed and you have VERY rich people that stand to lose a lot.

DeepSeek v3 is something I would actually install a VPN for, to get access to it, if it becomes an issue in Canada. It truly is on par with Sonnet and cost 15 times less.

My hope is, people will start to use DeepSeek v3 to do what DeepSeek is accused of having done and by the time things make it to court, it would be too late as there would be no way to say where any of the data came from.

Did you read the article? The amount of data taken was enough to trip wires.

Microsoft’s security researchers in the fall observed individuals they believe may be linked to DeepSeek exfiltrating a large amount of data using the OpenAI application programming interface, or API.

The article says linked. It could be someone's grandma living outside of China, or it could be a company outside of China with links to DeepSeek.

I suspect the goal is to make it not viable for any company to host the model and provide API services and create a choke point with deepseek.com

This is actually quite serious in my opinion, as running the 671B model is not trivial and I was hoping more vendors would pop up to provide API services.

Who knows with the Trump administration and their stacked Supreme court. I'm guessing what is going to be debated is not so much the data itself, but the means of obtaining the data. Basically there was breach of contract somewhere.

Based on the article (see archive link by bushido), this might actually get very serious. Looks like politician are now involved and I'm sure Altman and Ellison are going to lean hard.

---

David Sacks, President Donald Trump’s artificial intelligence czar, said Tuesday there’s “substantial evidence” that DeepSeek leaned on the output of OpenAI’s models to help develop its own technology. In an interview with Fox News, Sacks described a technique called distillation whereby one AI model uses the outputs of another for training purposes to develop similar capabilities.

---

Full disclosure: I am writing a chat app that is designed for software development

> It's a difficult problem to solve, requiring new sets of books, courses etc.

I think new tooling built around LLMs that fits into our current software development lifecycle is going to make a big difference. I am experiencing firsthand how much more productive I am with LLM, and I think that in the future, we will start using "Can you review my conversation?" in the same way we use "Can you review my code?"

Where I believe LLMs are a real game changer is they make it a lot easier for us to consume information. For example, I am currently working on adding a Drag and Drop feature for my chat input box. If a junior developer is tasked with this, the senior developer can easily have the LLM generate a summary of their conversation like so:

https://beta.gitsense.com/?chat=d36e0282-4326-46cf-83b1-4207...

At this point, the senior developer can see if anything is missed; if desired, they can fork the conversation to ask the LLM questions like "Was this asked?" or "Was this mentioned?"

And once everybody is happy, you can have the LLM generate a PR title and message like so:

https://beta.gitsense.com/?chat=8aa19528-5891-4dda-9a88-247a...

All of this took me about 10 minutes, which would have taken me an hour or maybe more without LLMs.

And from here, you are now ready to think about coding with or without LLM.

I think with proper tooling, we might be able to accelerate the learning process for junior developers as we now have an intermediate layer that can better articulate the senior developers' thoughts. If the junior developer is too embarrassed to ask for clarification on why the senior developer said what they did, they can easily ask the LLM to explain.

The issue right now is that we are so focused on the moon shots for LLM, but the simple fact is that we don't need it for coding if we don't want to. We can use it in a better way to communicate and gather requirements, which will go a long way to writing better code faster.

It's not too bad given that it runs in your browser. I took your prompt and asked GPT-4o mini to elaborate on it and got this https://imgur.com/a/qmQ7ZHl

The burger looks good.

> Sure sure that censorship is a problem, but that's a political background everyone knows, while none of the researchers of deepseek can do much about it, and literally do people think Chinese people like to put more efforts to censor LLM output?

I think it is a knee-jerk reaction without understanding how LLMs work. The beauty of all of this is, we can use DeepSeek and still give the CCP the middle finger. I don't know why people don't realize we can easily add a layer above DeepSeek to never ask it for political/historical information and we can easily create services to do this.

We should be celebrating what is happening as this might force OpenAI and Anthropic to lower prices. DeepSeek is FAR from perfect and it would be stupid to not continue relying on other models, and if DeepSeek can force a price change, I'm all for it.

They would, but I think challenging facts will be easier. Instead of saying "I heard if from blah" and not having an easy way to fact check it, you just get LLMs to challenge one another. LLMs (today's, not future versions which could be drastically different) don't have a built in goal post mover.

I actually think changing history will be harder in the future as it requires alignment across models.

I wrote a chat app built around mistrust for LLM responses. You can see an example here:

https://beta.gitsense.com/?chat=ed907b02-4f03-477f-a5e4-ce9a...

If you click on the Evaluation links, you can see how you can use multiple LLMs to validate LLM response. The evaluation of the accurate response is interesting since Llama 3.3 was the most critical.

https://beta.gitsense.com/?chat=fdfb053d-f0e2-4346-bdfc-7305...

At this point, you would ask Llama to explain why the response was not 100% which you can use to cross reference other LLMs or to do your own research.

> Anthropic have hit their maximum capacity

Yeah. They won't reset my API limit until February even though I have 50 dollars in funds that they can take from me. It looks like I may need to look at using Amazon instead.

> saying apple is dead because back in 1987 there was a cheaper and faster PC offshore

What Apple did was build a luxury brand and I don't see that happening with LLMs. When it comes to luxury, you really can't compete with price.

> But I don't think R1 is terminal for them.

I hope not, as I we need more competition.

> OpenAI has 100s of millions of daily active users, with huge revenues.

My rational is we are dealing with a commodity product. People will go where the best answer is. I only use DeepSeek because it is good. If it was free, but sucked, I would not use it.

Honestly, I do hope they (OpenAI) offer a better product but as it currently stands, I will not use their models because they don't offer enough value for the price.

I literally cannot see how OpenAI and Anthropic can justify their valuation given DeepSeek. In business, if you can provide twice the value at half the price, you will destroy the incumbent.

Right now, DeepSeek is destroying on price and provides somewhat equivalent value compared to Sonnet. I still believe Sonnet is better, but I don't think it is 10 times better.

Something else that DeepSeek can do, which I am not saying they are/will, is they could train on questionable material like stolen source code and other things that would land you in deep shit in other countries. DeepSeek just needs to improve the value and I can see them destroying Anthropic since I believe coding is their main focus.

When it comes to text processing, I personally find GPT to be much better and that might also have to do with allegations that they trained on literature that they should not have.

Prices will increase by five times in February, but it will still be extremely cheap compared to Sonnet. $15/million vs $1.10/million for output is a world of difference. There is no reason to stop using Sonnet, but I will probably only use it when DeepSeek goes into a tailspin or I need extra confidence in the responses.

Anything that becomes valuable will become a CCP property and it looks like DeepSeek may become that. The worry right now is that people feel using DeepSeek supports the CCP, just as using TikTok does. With LLMs we have static data that provides great control over what knowledge to extract from it.

Agreed. I am no fan of the CCP but I have no issue with using DeepSeek since I only need to use it for coding which it does quite well. I still believe Sonnet is better. DeepSeek also struggles when the context window gets big. This might be hardware though.

Having said that, DeepSeek is 10 times cheaper than Sonnet and better than GPT-4o for my use cases. Models are a commodity product and it is easy enough to add a layer above them to only use them for technical questions.

If my usage can help v4, I am all for it as I know it is going to help everyone and not just the CCP. Should they stop publishing the weights and models, v3 can still take you quite far.

> In other words: you must already know how to do what you are asking the LLM to do.

Those that will benefit the most will be senior developers. They might not know the exact problem or language, but they should know enough to guide the LLM.

> In other words: it may make sense if typing speed is your bottleneck and you are dealing with repetitive tasks that have well been solved many times (i.e., you want an advanced autocomplete).

I definitely use a LLM as a typist and I love it. I've come to a point now where I mentally ask myself, "Will it take more time to do it myself or to explain it?" Another factor is cost, as you can rack up a bill pretty quickly with Claude Sonnet if you ask it to generate a lot of code.

But honestly, what I love about integrating LLM into my workflow is, I'm better able to capture and summarize my thought process. I've also found LLMs can better articulate my thoughts most of the time. If you know how to prompt a LLM, it almost feels like you are working with a knowledgeable colleague.

> I never actually used it myself, I only gave into a suggestion by a friend (whom LLM reportedly helps) to use his LLM wrangling skills in a thorny case.

LLMs are definitely not for everyone, but I personally cannot see myself coding without LLMs now. Just asking for variable name suggestions is pretty useful. Or describing something vague and having it properly articulate my thoughts is amazing. I think we like to believe what we do is rather unique, but I think a lot of things that we need to do have already been done. Whether it is in the training data is another thing, though.