Видео

This was one of the questions I couldn’t find answers to. Great video Jason

please please please make more content we need competent ppl to make content cuz these marketers are flooding my feed with BS

There are so many glitches happening at once, please

More please!!!

Excellent office hour. Thanks for sharing

Such a great video - thanks Sam

So much importance rides on the quality of the data even though the hand labeling is tedious! Love it.

What happened to her right eye?

Nice intro, and good way to share the rag course content possibly. Thanks Jason.

Nothing beats data

Here are the top 10 takeaways: 1. Consultants vs. Contractors: Many people call themselves consultants but are essentially contractors, selling their time instead of outcomes. Successful consultants focus on outcomes rather than trading time for money. 2. Outcome-Based Selling: The key to scaling consulting is to move away from task-based selling to outcome-based pricing. Clients should pay for results, not tasks or hours worked. 3. Effective Landing Pages:A consultant's website or landing page should not focus on the consultant's skills but on the pain points of potential clients and how those can be resolved. It’s about framing problems and showing the outcomes. 4. Understanding Client Pain Points: To be successful, consultants need to identify and clearly articulate the specific pain points of their clients. This helps in crafting proposals that are relevant and outcomes-focused. 5. Multiple Pricing Options: Always offer multiple pricing options in a proposal. Each option should reflect different levels of commitment and accountability for the client’s outcome. This gives clients a choice and anchors higher pricing tiers. 6. Value-Based Pricing: Instead of attaching costs to time or tasks, structure pricing around the value delivered. Higher prices should reflect greater accountability and responsibility for delivering results. 7. Free Calls and Content: Jumping on free calls and producing content related to client problems is an effective strategy for gaining trust and positioning oneself as an expert. The content can serve as a sales flywheel. 8. Metrics and Accountability: To add value, it’s important to include metrics that show improvement and success. This helps clients see the impact of the work and builds trust in the consultant's ability to deliver. 9. Test and Iterate on Proposals: Proposals should not be static. Iterate and learn from them, offering tiers that provide different levels of service and accountability, and adjusting based on feedback and results. 10. Proof of Expertise: Building a library of case studies, testimonials, and content that proves past successes is crucial for long-term client trust and acquisition. Showing proof of concept and expertise through real-world examples drives conversions.

Would you change your discovery strategy on outbound vs inbound? It makes sense how you are able to ask all of those questions on quantifying the pains because you already know they want to work with you. If someone had no idea who you were and had never read your blogs or twitter, how would you approach that conversation? This content was awesome! Thanks

Interesting. From my experience with annotators, I found it was less about "firing people" who performed bad and perhaps more about "rewriting the guidelines". Sometimes an annotator takes the guidelines literally (actually not a bad thing) and as a result generates annotations that the guideline designer did not have in mind. This is also partially why it makes a tonne of sense for folks who write guidelines to also annotate on the task. It can also help to have an annotation interface where folks are able to flag a task/example as confusion so that it's easy to reflect. I have not tried it with LLMs, but my gut says that allowing the LLM to flag an example/task combo as confusing can also really help in designing a few solid prompts.

love this kind of video! learned so much

Jason, how come the your free content on yt etc is aimed at other consultants (competitors) and not your actual customers? It seems like a missed marketing opportunity unless your customers are actually AI consultants and not companies in industry?

If you think services aren't a great model, why don't you go full-time into SaaS?

This is great

man, this is a goldmine, thanks!

Great discussion, thanks.

Can you please add links to references (like Eugene's post) in the description? Great content!

Thanks, this was really good.

Great point. I'd love to hear examples of different decision-making reports in a future video!

So good!

Thanks, this really helps!

We have built exactly this. Happy to see someone who aligns with our hypothesis :)

Jason, excellent point on the higher level output for a RAG. Out of curiosity, what was the book you mentioned on sales engagement letters?

Interesting take on the report generation because of the obvious value for a company. Curious to hear from you how this would defer from using standard calls to a long-context LLM, if and when they get to a point of desent latency and long-retrieval understanding.

I thought someone was about to grab you from behind in the thumbnail 😭

turning the ambiguity of language and communication into concrete and well structured insights that can drive effective business decisions with the help of LLM's ?

noticed this as well - general rag without domain expertise is lackluster. Using llm calls as a compute primitive to build extremely good verticalized software, much better.

repost?

I agree, but would replace "report" with "artifact" as a general output category.

Thanks for making Instrucor bro

Where can I find code for this

Great converstation and insights guys!

I just found Instructor. Cannot thank you enough! This is the most elegant way to program LLM interactions I've seen!

who are coaching you when it comes to consulting?

Its great to listen to your guys! I am a python / C# .Net guy, but the last 2 years been nothing but pure excitement learning everything i can about AI Agents and all the new design pattern and architetures that the LLM as the central magical black box. By the way your instructor class is wonderful... i have integrated that into my Agent Builder tools and its working great! thanks yall.

What’s rag ?🙏

slay

Jason thank you for dropping all this wisdom on us! This short video format on separate topics is great.

What do you think about Live-Updating RAG? Let's say some info changed in some client's watched doc, we would want to immediately prepare it for rag, remove the old outdated chunks, rechunk new data, and update the db/vdb. Ideally very quickly. I think it really sets apart fun-to-do projects and real world production systems. Would love it if you could cover this 💜 Thanks & All the best! ps - Your takes around benchmarking metrics are spot on. We need to develop better tooling for ourselves

Is there any reference implementation on how to address routing/query classification? considering multi-turn, context aware conversations.

Hey Jason, great takes! This whole thing almost sounds like trying to develop a timezones system from scratch haha. I'm guessing very soon RAG will be "solved" (sota foss) but we're still at the bleeding edge. I think most of the meta "non embeddings" questions you're asking could be answered by a small script over the db, if the llm knows the structure of you data and the right keys to search, it could write the extraction code and provide an answer with 2 steps. Also for the most common questions, the answers could be ragged and live updated

Im not exactly sure where you were going with it either lol but i definitely agree with some things you said

Very useful concepts to think about. Thank you for the video.

This is so enlightened me, Thanks! ❤

Hey Jason! I've had a lot of success using instructor with open ai models, but not much when using local models (such llama3 and Mi(s/x)tral via ollama for example). Any suggestions on how to get those to work well with structured data? Thanks!

Amazing advice, all of these things struck internal conversations I have had

Absolutely loved this one. 🫡