mvkel

17 ms·

301.
by mvkel 11mo ago
Predicting the stock market will likely never happen because it’s recursive. We can predict the next 10 days of weather, but the weather doesn’t change because it read your forecast. As long as markets continue to react to their own reactio
302.
by mvkel 11mo ago
S&P 500 is also tech heavy and notoriously difficult to beat over the long run
303.
by mvkel 11mo ago
When the market is rising, everyone looks like a genius. Would have been better to have variants of each, locked to specific industries. It also sounds like they were -forced- to make trades every day. Why? deciding not to trade is a good
304.
by mvkel 11mo ago
This is what makes Teslas sustainable and other car cos, like Porsche, not. A battery pack for a Model 3 is $10K. So even if the whole car is only worth $20K, it's still worth keeping on the road. The Porsche Taycan battery pack is $70
305.
by mvkel 11mo ago
Agree with all of these. Remember playing tag with your friends at recess? That's what exercise should feel like.
306.
by mvkel 11mo ago
But you can't remove something from the internet once it's there, so once it's released, it's expected that it always will be. It's also impossible to guarantee a 100% secure infrastructure, no matter how good your
307.
by mvkel 11mo ago
That depends more on what the Privacy Policy is of the service, which you agree to when you sign up and use it
308.
by mvkel 11mo ago
It does. Most privacy laws are based on time-from-discovery. If they immediately sprung into action at the moment they were informed and remediated the issue, they're in compliance.
309.
by mvkel 11mo ago
It's not a rumor, it's confirmed by OpenAI. All "models" since 4o are actually just optimizations in prompting and a new routing engine. The actual -model- you are using with 5.1 is 4. Nothing has been pre-trained from s
310.
by mvkel 11mo ago
I continue to be surprised that the supposed bastion of "safe" AI, anthropic, has a record of being the least-open AI company
311.
by mvkel 11mo ago
Open weight LLMs aren't supposed to "beat" closed models, and they never will. That isn’t their purpose. Their value is as a structural check on the power of proprietary systems; they guarantee a competitive floor. They’re es
312.
by mvkel 11mo ago
I think the protections are smarter than the author realizes. The reason it allows the image to be generated is because the output is clearly -not- in Elon's likeness. It looks like a crude photoshop of Elon's head pasted onto som
313.
by mvkel 11mo ago
I was excited to see this, but then realized only chapter 1 is done out of what ultimately will/should be a 25 chapter tome. Strudel docs leave something to be desired as well. What I've found to be the most useful so far is to as
314.
by mvkel 11mo ago
DC
315.
by mvkel 11mo ago
The problem with time-bounding is it becomes an obsession with meeting optimization, and the content of the meeting comes secondary. "Ok let's quickly do X Y and Z so the meeting can be over and we can be outta here!" "H
316.
by mvkel 11mo ago
There's an interesting housing arbitrage happening right now. Townhomes are still selling for over asking, but single family homes -- ones that older, middle-management types would live in -- aren't receiving competitive bids.
317.
by mvkel 11mo ago
Worth noting: this paper was published three days before the release of GPT-3.5
318.
by mvkel 11mo ago
Documenting for AI exactly like you would document for a human is ignoring how these tools work
319.
by mvkel 11mo ago
Work backwards from the customer. If you want to generate revenue, you'll need to work with b2b saas companies that can afford your product. That means they're doing (minimum) $500K ARR; a proxy for having some semblance of a team
320.
by mvkel 11mo ago
When AI starts topping the music charts, getting the biggest laughs, or writing the stories we prefer, the only thing it should be for creators is a mandate to do better
321.
by mvkel 11mo ago
What a cool piece/person/perspective. We need more of this unorthodox thinking in the zeitgeist
322.
by mvkel 11mo ago
I don't think it's this binary. Influence and power are far more intoxicating currencies than affiliate revenue. And if someone complained "you're just publishing this helpful thing to become more influential in [commun
323.
by mvkel 11mo ago
How could we hope to solve the alignment issue in AI when we're so terribly bad at it as a society?
324.
by mvkel 11mo ago
> One guy on the Internet is--and always will be--an anecdote It seems like you have a specific person in mind as the audience member (yourself?), but the piece could benefit from a wider view. Given your hardware choices, reliability s
325.
by mvkel 11mo ago
Waiting for the edit to the edit
326.
by mvkel 11mo ago
Wait. You build a new one every -year-?! How does one establish the reliability of the hardware (particularly the aliexpress motherboard), not to mention data retention, if its maximum life expectancy is 365 days?
327.
by mvkel 11mo ago
This piece feels a lot like someone criticizing an umpire's call after watching the slo-mo fifteen times and concluding the ball was actually a strike. Way different from the umpire's pov
328.
by mvkel 11mo ago
Open models serve as a safeguard, not a competitor, to closed models. They still cost billions to pre-train
329.
by mvkel 11mo ago
Microsoft, largest software company in the world, has (publicly and privately) admitted that it is in their best interest to ensure that OpenAI remains the leading ai company for at least the next seven years. As for nvidia, if OpenAI has l
330.
by mvkel 11mo ago
OpenAI is also very profitable if they deprioritize... scaling up.

More ›