tkunze

Check it out on afterflag.com

Why add a AI observability platform to a @Firebase app when Firebase Storage can hold agent traces? So I made AgentPond use it: $ npx agentpond init Own your traces. Analyze them locally. github.com/marcusschiesser/a…

$META CEO Mark Zuckerberg: “There’s a very high chance the world is going to look pretty different in a few years. We tell our teams internally: ‘Take superintelligence seriously.’ We believe this will reshape all of our systems—not necessarily quarter-to-quarter, but over the next few years, it’s going to change a lot of assumptions around how things work.” “Some of the most aggressive assumptions about AI timelines have turned out to be the most accurate. That’s why we’re pushing hard on talent, compute, and deploying AI across all our products—because Meta is uniquely good at getting technology in front of billions of people.”

Of course that's your contention. You're a VC associate turned wildfire expert, just got done reading some American Dynamism, probably @pmarca? Next, you'll be talking about the 2014 water bill, the smelt and blaming Newsom til you realize...

The Washington Post yesterday: the Ukrainian Army is exhausted, is waiting for a cease-fire, and has been set for defensive actions. The BBC today: the Ukrainian Army has launched a new Kursk offensive. Me, freshly from Sudzha: silently smiling.

.@AStuttaford

Hard to understand why 8 years after Trump's 1st election victory Europeans are still so reliant on the US when it comes to defense. They've really put their national security in the hands of swing voters in a handful of US battleground states. And it looks like that gamble has not paid off.

The two-state solution nobody is considering.

System Initiative is the future and it's ready for you to replace Infrastructure as Code. Read more about why on our blog, systeminit.com/blog-system-i…, and get started today: systeminit.com youtu.be/20vIl7I0R4I

Donald Trump's 'Eating The Dogs' transcribed for trombone.

5 week glow up period post breakup:

Loved this conversation between @dwarkesh_sp, @fchollet and @mikeknoop! youtu.be/UakqL6Pj9xo Highly recommend listening to the full episode, but here are a few interesting notes and thoughts: (1) It’s fascinating to see how much experts in AI/ML disagree not only about the future (e.g. how much room there is for further scaling of LLMs, Hinton/Altman/Sutskever vs LeCun, Gary Marcus etc) but also about the capabilities of LLMs *today*. Dwarkesh shared an anecdote of talking to experienced ML researchers (who didn’t know about ARC) and showing them an example ARC problem. They said, 'of course, an LLM would be able to solve something like this’. They took a screenshot of it and put it into ChatGPT – it didn’t get the pattern. (2) François argues that LLMs are not intelligent. No matter how intelligent e.g. ChatGPT appears to be – it’s all just pattern-matching and memorization. If you look at some of the remarkable achievements of ChatGPT (e.g. interpreting an essay that it has never seen before; learning a language based on a grammar book and a dictionary alone; …), it’s hard to imagine how it can do this without intelligence. However, the size of these models is so far beyond what we can grasp that we can’t trust our imagination here. The counter-argument is that humans (in general, not only VCs 🤣) rely on a lot of pattern-matching, too. Here’s where ARC comes in: These puzzles are designed to be easy for humans, but hard for AI (see arcprize.org). (3) Another argument against François’ position might be that it’s anthropocentric. Machine intelligence will look very different from human intelligence, so let’s not theorize too much and just look at the results. If a computer can do lots of tasks that previously required human intelligence … well, then I guess the computer is intelligent. Let’s give the machines the benefit of the doubt instead of continuously moving the goalpost. ;-) A corollary of the tight definition of intelligence is that if or when a machine can do the job of, say, a junior lawyer or junior developer, while not being intelligent based on François’ reasoning, that job apparently doesn’t require intelligence. (4) It may seem like this is all just semantics, but there are some real implications because the level (or absence) of intelligence of LLMs will tell us something about the speed of future progress in AI (obviously a hot topic in SaaS these days). It should also have an impact on how much research should go into other AI areas outside of LLMs. Quoting François: “OpenAI basically set back progress towards AGI by quite a few years, probably like 5-10 years. […] They also triggered this initial burst of hype around LLMs. Now LLMs have sucked the oxygen out of the room. Everyone is just doing LLMs. I see LLMs as more of an off-ramp on the path to AGI actually. All these new resources are actually going to LLMs instead of everything else they could be going to. If you look further into the past to like 2015 or 2016, there were like a thousand times fewer people doing AI back then. Yet the rate of progress was higher because people were exploring more directions.” (I don’t agree with this but it’s a refreshing perspective and great food for thought.)

People cannot accept that we are pretty predictable. It's far more comfortable to pretend we've created a digital god than it is to acknowledge how formulaic most human output actually is.

Inflight wifi didn't work so of course I had to debug it. It appears the problem is lack of DHCP lease. The WiFi was using 8 hour leases, which was time enough for many planeloads of passengers to embark/disembark. A quick ARP scan at the time showed there were 55 devices on the network, almost entirely mobile devices (which randomize their MAC addresses, whereas laptops don't). Given the number of takeoffs and landings in the 8 hour period, the lease table can easily fill up. The fun fact about DHCP is that when this condition happens, it sends no response to the client. In other words, when trying to diagnose why " the wifi isn't working", there's no difference between this cause and half a dozen other causes. I have to guess this is the cause simply by looking at the fact lots of other people seem to have successfully gotten a DHCP response but I haven't. I told the flight attendant "the WiFi isn't working". She then "reset the Internet", after which I could get a lease. Apparently in the front near the entrance/exit, there's a button simply labeled "INTERNET RESET" that she presses whenever a customer complains. Obviously, one solution to the problem is that DHCP leases on planes should be drastically shorter, like at 1 hour intervals. Secondly, the number of leases should be drastically increased.

who can relate? 😂