Big news for a clean planet.
Bill Gurley
628 items collected
628 matching items · newest first
State by state regulation in any field creates massive bureaucracy & inefficiency. Doing it now in AI would drastically hinder US efforts to be competitive worldwide. It was dumb in CA (1047) and it’s equally dumb in every state.
@NathanCHubbard Maybe a fresh ink job.
This was always my biggest fear about NIL. 99% of student athletes didn’t have an issue, but we changed the system because we thought the 1% were not being treated fairly. Now the 99% (all genders in all NCAA sports) may suffer so that a few judges and elites feel better about how the 1% are treated. And keep in mind that age limits in NBA and NFL contributed to this problem. No one worried about golf or tennis or even baseball. You could just go pro.
Texas is making changes to further separate it from the pack as many companies smartly look to exit Delaware.
PAGA is horrific. Every lawyer is deputized. Lawyers also get paid.
@mario_cibelli But you are recreating options.
@mario_cibelli Metoo @mario_cibelli Zero basis time vested RSUs do something options never did. They pay exec bonuses for zero value creation and even payout with value destruction. It’s unreal that ISS led us here.
If on a board you are “supportive” or even “suggestive” that it’s a good idea to buy back stock to offset SBC dilution, you are confirming what was already true. SBC is an exact equivalent to spending cash on OPEX. Adjusted EBITDA that subtracts SBC is financially ignorant.
@ENERGY Yes. Fission. Typo.
Did @Jason arrange this?
Important to have clear eyes. https://t.co/RZs6lS5jEq
I have been thinking about @mattwridley’s book quite a bit while studying LLM evolution. Especially when a model can train or learn from another model. Now, what if in a single country there were four open models that could all learn from each other?
The math doesn’t work. Worth a listen.
I warned about this 12 years ago. https://t.co/vLHtDC632Q
Based on the rumor mill, it sounds like the closed-model US AI company’s lobbying efforts were having some success persuading the US government to ban or restrict Deepseek. Now comes Qwen. Then MiMo. Open comes at you fast. What now?
China loves the open source.
I agree with this. Probably better for the government to figure out why our fission nuclear is 4x the cost of China and South Korea than worrying about NVidia or banning Deepseek.
I truly believe the idea of "winning the AI war" is misplaced, unachievable, and will lead to horrible and ineffective policy. More below ...
Super important reality highlighted here by @joshk, but I fear the real-life experience is even worse. With most of the contemporary equity incentive programs, companies experience 3-6% dilution annual EVEN IF the value of the company is flat or down. So time hits twice.
I agree with this 100%. I blame the large about of energy put into EBITDA based multiples. Obviously you can pay people w stock that is immediately sellable. That is same as cash but added up differently when we talk “adjusted EBITDA”. This is just weak finance.
Same.
“It's time to move on, time to get going.”
Steph doing Steph things…
@htifc I think you are right.
Must watch: @dee_bosa hits this in the center of the target. Our nation's recent curbs on NVidia H20s, intended to slow China AI innovation, will enhance & accelerate Chinese AI innovation. This action will create the OPPOSITE of the intent. https://t.co/olhHjl1aCC
🤫 https://t.co/25gemZYS8O
Important and timely conversation between @bariweiss and @nfergus. Must listen analysis of current tariff policies.
Peer respect is one of my favorite things. 😎🤠❤️
@AishwaryaNaga Have to believe auth granting will be a thing. Seems dangerous. Which company do you trust?
Collection coverage
facebook: unchecked
linkedin: unchecked
podcast: unchecked
reddit: unchecked
substack: unchecked
web: linked
x: linked
youtube: unchecked