Derek Andersen (@DerekjAndersen) header
Derek Andersen (@DerekjAndersen) profile picture

Host @Divotdotorg, co-founder @StartupGrind/@bevyhq Labs. Posting about: tech, startups, AI, community, Silicon Valley, 4 kids, @BYU, Believer. Still grinding.

Palo Alto, CA
Joined February 2008
Photo by Derek Andersen (@DerekjAndersen)Photo by Derek Andersen (@DerekjAndersen)Photo by Derek Andersen (@DerekjAndersen)Photo by Derek Andersen (@DerekjAndersen)Photo by Derek Andersen (@DerekjAndersen)Photo by Derek Andersen (@DerekjAndersen)Photo by Derek Andersen (@DerekjAndersen)Photo by Derek Andersen (@DerekjAndersen)Photo by Derek Andersen (@DerekjAndersen)Photo by Derek Andersen (@DerekjAndersen)
The hardest lesson NFL MVP Steve Young ever had to learn. "Fear-based thinking is a liar. I was afraid of failure, so much so that I paid this price so that I could have success. I learned as an older person, that's a fallacy. That's fraudulent." "I'm not going to pay for success with my worry and fretting. Man, fame is a thief. Fear-based thinking is a thief. There's a lot of thieves in life."
47
449
50
4,170
782,974
Just remember, you can’t complain about politics if you don’t vote! Or in California if you just show up and pick a random name and address and say it’s you, you can vote for them. 10x higher security to get into my kids’ homecoming dance.🤦‍♂️🤦‍♂️
6
3
87
2,711
The fear of rejection costs more than rejection itself. James Dumoulin (@hardknocksedu ) on how he approached Mark Cuban and landed his first billionaire interview: “I'm a firm believer that the world is overly cruel to shy men, and it rewards the courageous.” “But then there's something that I live by, which is this idea of take the risk or lose the chance.” “And he happily said yes.”
2
3
22
1,311
We were born to win in life, and winning starts with belief. Brian Santiago on his competitive drive and the lesson his dad taught him: “We were all born to win, not just in sport, in life.” “I just felt like I could find a way to win.” “…he didn't want me to ever lose my competitive edge, but that I needed to channel it…”
2
3
1
51
5,275
Happy Halloween!
15
869
Waymo completely bricks on Halloween street. Living in a real life episode of Silicon Valley.
14
1,759
Best place to start Halloween day. Have a great and safe day!
2
33
2,013
Currently in college application essay hell. Thoughts and prayers please.
14
1
81
5,206
Zuckerberg’s open AI model was like a $100 million loan. Blackbox AI co-founder Robert Rizk recalls an early investor’s reaction to Meta’s Llama: “And when Llama launched, he told us during a meeting that Zuckerberg just gave you a $100 million loan.” “Because as developers, you needed to have access to somehow competitive models, and that was one of the big drivers.”
5
12
1,580
Best reason to walk to B gates at SLC airport: Swig.
1
15
796
The AI race will be won by whoever gets the most from every GPU. @blackboxai co-founder @RobRizk1 on what will separate good AI companies from great ones: “The difference between good and great companies is how do you optimize inference so that actually you can produce more tokens per GPU.” “Inference is going to be powering everything, every single product that you interact with.”
5
6
16
997
We can all relate to being backed into a corner in a desperate situation and Red Baron Pizza guy's first reaction was to kindly help. GoFundMe already raised $65k. Donate if you can. gofundme.com/f/support-a-har… Hey @redbaronpizza: why don't you match whatever is raised for him?
‘MY LIFE IS HARD AS IT IS, I DON’T NEED THIS CRAP!’ — Grocery store worker SNAPS at clout-chasing kids Teens tried to apologize to the man with four kids and a mortgage, but it didn’t look sincere to him
26
26
3
345
52,790
Rise and shout…fun interview today with @ddosanjhtaylor coming soon.
4
1
64
3,970
Derek Andersen retweeted
Replying to @DerekjAndersen
@DerekjAndersen is one of the few people that had significant impact on where we are today as a company!
Inference may be the biggest AI trend of 2027. @blackboxai is an emerging Inference leader, having grown revenue from $0 to $300M in the past year. Their platform is winning because it’s end-to -end encrypted, totally US-based (no data sent to China), and in many cases twice as fast. My interview with the founders below. Timestamps: 01:05 — AI Inference: Why More Tokens per GPU Matter 04:08 — Why Businesses Want Both Open and Closed AI 09:54 — DeepSeek, China, and Where Your Data Goes 18:43 — Encryption vs. Zero Data Retention 25:40 — Lowering AI Costs With the Same NVIDIA GPUs 31:06 — Why Smarter Models Still Need Better Infrastructure 38:01 — Three Brothers, One Company: The Blackbox Story 43:14 — How Trust and Disagreement Help Founders Move Faster 47:00 — The Navy SEAL Principles Behind Blackbox’s Culture 01:01:28 — Overcoming Hard Problems, One at a Time #AI #AIInference #BlackboxAI #DivotPodcast
2
1
10
524
Inference may be the biggest AI trend of 2027. @blackboxai is an emerging Inference leader, having grown revenue from $0 to $300M in the past year. Their platform is winning because it’s end-to -end encrypted, totally US-based (no data sent to China), and in many cases twice as fast. My interview with the founders below. Timestamps: 01:05 — AI Inference: Why More Tokens per GPU Matter 04:08 — Why Businesses Want Both Open and Closed AI 09:54 — DeepSeek, China, and Where Your Data Goes 18:43 — Encryption vs. Zero Data Retention 25:40 — Lowering AI Costs With the Same NVIDIA GPUs 31:06 — Why Smarter Models Still Need Better Infrastructure 38:01 — Three Brothers, One Company: The Blackbox Story 43:14 — How Trust and Disagreement Help Founders Move Faster 47:00 — The Navy SEAL Principles Behind Blackbox’s Culture 01:01:28 — Overcoming Hard Problems, One at a Time #AI #AIInference #BlackboxAI #DivotPodcast
3
6
1
21
1,918
AI agents can create parallel worlds to solve one problem. @daytonaio co-founder @ivanburazin on how agents can explore multiple paths in parallel: “Agents can, theoretically, for anyone that's watching the Marvel movies, go through the multiverse.” “So every decision it can make, you can essentially replicate the computer and try both of the outcomes. And then you can replicate it again.” “So for every fork in the road, you can basically split that up.”
4
3
15
1,580