Tremendous

An angel investor's take on life and business

DeepSeek just dropped a new model, R1-0528. Has China leapfrogged the US? This morning, I tested the model to find out.

DeepSeek shocked the world with its R1 release in January. But when I put it up against the best from Gemini, DeepSeek lost.

Let’s see how DeepSeek’s latest release performs…

Timing My Workouts

I need to get a brief HIIT bike workout in this morning. But I’m not sure of the best time to do it.

I ate a light breakfast about an hour ago — is it too soon to work out? Let’s see what DeepSeek says…

DeepSeek gives me a variety of options in a nice table format. My breakfast was on the larger side in the context of their table, although I’d still consider it light for me. I’m an eater! 🙂

So, it sounds like I need to wait around 2 hours. Guess I’ll blog instead!

This was a pretty good response, although I wish they had quantified the breakfasts in terms of calories. I’d also like to see some citations to experts on exercise.

Overall, I’m giving this answer a B.

SaaS Companies Making It Big

In 4 years investing, I’ve had a few companies go bust. But luckily, I also have a few that are beginning to break out.

Two have cracked $20 million ARR so far. But how rare is that, really?

Let’s ask DeepSeek how many companies make it to that level…

DeepSeek estimates that only 1 out of every 200-500 SaaS startups ever cracks $20 million.

DeepSeek gives some citations, but there’s no clickable link to verify the info myself. That makes me wonder if it’s even real.

So, I tried to verify the data on my own. And guess what that involves?

A trip to Google! I may as well have not even bothered with DeepSeek.

This query demands a data driven response. The whole question involves statistics. But DeepSeek couldn’t get it done.

I’m giving DeepSeek a C here.

Startup Research

This evening, I’m meeting with a company that’s using robots to automate scientific research. I’ve never seen anything like it.

But I wondered, is anyone else doing this? Let’s ask DeepSeek…

DeepSeek gave a pretty strong answer here! It turns out that although automated, robotic research is new to me, it’s not exactly new.

A company called Strateos is doing it, as are several major drug manufacturers. I’ll look into their platforms more this afternoon before my meeting.

I would’ve liked to see some citations here. Instead, I have to go hunting on my own.

But overall, DeepSeek gave me a pretty good answer. I’ll give this round a B+.

Wrap-Up

DeepSeek’s answers would’ve been impressive 2 years ago. Not anymore.

The responses were informative and seemed reasonably accurate. But without any citations, it’s hard to know how accurate.

I can get similar responses from many models. Llama, Claude, Mistral, they could all easily do what DeepSeek just did here.

DeepSeek is behind the leaders, OpenAI, xAI, and Google. Those products do much deeper research with better citations.

Overall, I’m giving their new model a B-.

So far, DeepSeek is mostly hype.

More on tech:

DeepSeek vs. Gemini Deep Research: Which Model Is King?

How Grok Saved Me from Getting My Identity Stolen

Google Glasses: A Computer for Your Face?

Save Money on Stuff I Use:

Fundrise

This platform lets me diversify my real estate investments so I’m not too exposed to any one market. I’ve invested since 2018 with great returns.

More on Fundrise in this post.

If you decide to invest in Fundrise, you can use this link to get $100 in free bonus shares!

Misfits Market

I’ve used Misfits for years, and it never disappoints! Every fruit and vegetable is organic, super fresh, and packed with flavor!

I wrote a detailed review of Misfits here.

Use this link to sign up and you’ll save $15 on your first order. 

2 responses to “Testing DeepSeek’s New Model”

  1. […] DeepSeek beat the best American models? When I tested DeepSeek in June, the outputs were garbage. This morning, I gave it a […]

    Like

Leave a reply to Amazon’s Nova Model: Cheap and Underrated – Tremendous Cancel reply