What? No. Making the same mistake repeatedly is the most predictable thing in the world. Until he gets confirmation that he's shooting down the wrong side, he thinks he's just racking up a great kill streak of Migs.
I thought so too... but if you refresh the page, it's just a pre-baked animation. A fun idea for somebody though; a little aquarium full of bots doing fake office tasks (I'm sure it's been done already).
The animations are CSS-driven, but the data behind them is real — agent heartbeats, task counts, and activity logs are pulled from actual systemd timer outputs. It's not a mock dashboard, though I'll admit the visual polish probably makes it look more "produced" than a typical monitoring tool.
"Fake it till you make it" in a nutshell. Half the AI wrappers on the market do the exact same thing: they render pretty activity charts that have absolutely zero correlation with actual VRAM consumption or server-side inference latency
Guilty as charged on this one. The dashboard visuals are CSS animations — the data behind them isn't live yet. I've been trying to pipe real systemd logs into it but haven't cracked the architecture cleanly enough to ship it. It's on my list, just not done.
Should've been clearer about that in the post. Thanks for pushing on it.
Thanks! The dashboard is one of my favorite parts of this project. It actually pulls real agent activity data — the visualizations are live, not pre-rendered. Working on making it a product: customizable agent dashboards for other teams running multi-agent setups.
Very good question - posted it for awareness / sparking hopefully nuanced “are we the baddies here?” reflection in the community, and curious folks to preorder.
If anyone ever writes a post of why that error keeps happening with browsers that should support it, I'd be incredibly grateful. Keep seeing it in our (unrelated to OP company) Sentry logs and zero chance to reproduce them.
You still have a lot of credibility to not be put into the number-go-up bracket and the social capital to overcome the political and power structures you had to face for two decades and know more about than most people by having built your company.
But as long as I don't see somewhat more transparent conversations with the people in your orbit like patio11, Matt Levine, Kyla etc, where you address how you'll actually tackle the non-technical challenges ahead, this GTM communication and site looks like every other 2019 JPM, HSBC etc "something blockchain" announcement and hard to get behind as something that might as well be really different this time, and not be killed/sidelined by vested interests. Including your own.
The fact that Stripe is doing it in 2025 should already be a strong signal. If they were just like the clueless trendmongers who ran crypto initiatives at large institutions in 2019, they'd have done it then instead of now, after the GENIUS act has been passed.
It’s a hilarious definition of “fix” but the basic argument is that when you mandate stablecoins to hold treasuries and they start seeing actual adoption, you create demand/sink for a couple extra trillion dollars of treasury bonds.
Eg if Australian locals suddenly switch transacting cocaine at scale in Tether instead of AUD, the US government can borrow more money by providing that collateral to Tether.
In this scenario, what happens when there’s a run on a stablecoin?
The bonds are sold en masse, and the value of those bonds will be hit, driving up gov borrowing costs (plus they just lost a source of demand), meaning the stablecoin “bank” could be bankrupt, right?
With stable coins you’re really trusting a private company to invest your money in a way that is robust to a drop in confidence. Isn’t this high risk? If a coin gets large enough, is it a threat to government solvency?
I’m sure the current administration will put prudent oversight in place for that not to happen.
But I guess we will find out in a 2027 Bessent presser announcing the Fed stepping in.
More serious answer: the bigger risk is trusting SV types to be content with a couple of percent in spread, and not starting to pull all kind of shenanigans to juice returns to a point where it becomes much harder to bail them out vs just taking back the treasuries.
US government solvency seems, as crazy as it sounds, less of an issue, as evidenced by the brief tantrums with absolutely no real effects beyond a couple of protesting headlines in the recent months. Where else are people around the world going to put their money? But as gifted as the current gov crew is at turning privilege into disaster, we're probably going to find out soon enough if there are any actual limits to that.
Normally not too much sympathy for the 7 or so 'sober' ECB folks, but I do not envy them for having to endure their hyped up CBDC colleagues at Robert Johnson at 3am…
Thank you for providing an actual answer! I don't know what to think of it--it seems naïve to say the least. But it is at least a self-consistent answer that helps me better understand what people mean when they say stuff like that.
Nice! Need to implement realtime lightning data in a project soon, WIS2 is great for overall weather details but doesn't have a good temporal lightning resolution. Has anyone reached out to both and done that recently with WWLLN and/or Blitzortung?
The former seems to have better coverage especially across the southern hemisphere.
Thanks a ton! Was afraid that that's the answer - and that there's no reasonably priced aggregator/abstraction layer, eg like https://open-meteo.com for ECMWF.
You rang ;-) I’m in the middle of adding more ECMWF data that will be released as open data starting October 1st. At the moment, only a limited set of lower-resolution (0.25°) ECMWF forecasts can be shared open-data. That’s going to change in a big way, though I can’t share more details just yet.
I’d still look into fine tuning. One of the main issues I’m struggling with is that even reasonably simple drivers often need to be aware of Application Notes etc that would single handedly blow past any context window. So giving the model a fighting chance by narrowing the space might be worthwhile here. Top 20 vendors datasheets/docs, handpicked sites and forums etc.
we are talking to a some of the larger vendors about getting first-party support for documentation. once we scale up a bit I think it could make sense to have multiple fine tuned models based on the docs they provide (or we crawl the web for public ones, at minimum)
