Half the cost per task compared to Opus 5, comparing high effort to high effort. That's just really nice.
Edit: https://artificialanalysis.ai/models/claude-opus-5-5?models=...
2,578 karma · joined October 28, 2009
previously cofounder of periscope data
Edit: https://artificialanalysis.ai/models/claude-opus-5-5?models=...
When I graduated in 2007, it was common for tech companies to refuse to let their systems be used for war, and it was an ordinary thing when some of my graduating classmates refused to work at companies that did let their systems be used for war. Those refusals were on moral grounds.
Now Anthropic wants to have two narrow exceptions, on pragmatic and not moral grounds. To do so, they have to couch it in language clarifying that they would love to support war, actually, except for these two narrow exceptions. And their careful word choice suggests that they are either navigating or expect to navigate significant blowback for asking for two narrow exceptions.
My, the world has changed.
Very weird that the foundational LLM companies' own chat pages don't do this.
"We were halfway through shooting season 1, coming through Covid, and the monumental size of the show, the effort, and everything else was just dawning on us. We realized that I didn't have enough calories to do it, and Diego's face couldn't take the timing, because it just takes too long to make it."
https://www.gamesradar.com/entertainment/star-wars-tv-shows/...
"By that point, the work that was required to make the show, at its minimum, was just dazzlingly blinding to look at. And Diego was like ‘Oh my god, we told them we’d do five years.’ Nobody, if we were gonna do it like this, you couldn’t physically do it. It was just impossible."
https://screenrant.com/andor-tony-gilroy-original-five-seaso...
uv is a clear improvement over pip and venv, for sure.
But I do everything in dev containers these days. Very few things get to install on my laptop itself outside a container. I've gotten so used to this that tools that uninstall/install packages on my box on the fly give me the heebie-jeebies.
Yeah, the hack you described for CIs is typical of "80% charting". We have a list of probably thousands of longtail visualization requests and we're way past the 80/20 point.
These days customers who want to go 100% use the Python/R editors and do their custom visualization there. So you do your SQL query like usual, but then pipe it to Python/R for the visualization. Have you tried that, and has it worked for you? Or do you prefer another model?
In a perfect world, how would the charting in Periscope work? How would you want to go from getting 80% very fast, to going to 100%, in the same software?
`jplitz and `ajones are right. As of now we support MySQL, Postgres, Amazon Redshift, Vertica, SQL Server, Oracle, MemSQL, Sybase, Exasol and Google BigQuery. We add more all the time.
Periscope Data supports cross-DB joins. We cache the data in our own clusters so queries run really fast, and also so you can do things like join across databases and upload CSVs.
More info: http://www.periscopedata.com. I'm harry at periscopedata.com
For ping-and-visualize, I'd recommend a tool that'll give you raw access to the underlying data. Some choices are https://amplitude.com/ and https://segment.com/. Happy to chat about pros and cons if you're curious -- harry@periscope.io.
There are two kinds of charts: Charts designed to find information, and charts designed to sell information. The latter are often gorgeous and many-dimensional: Heatmaps, animated bubble charts, charts with time sliders, etc. And by all means, if selling the data is required, then sell it with the best tool for the job.
As for actually investigating the data, it's usually a lot of tables, lines and bars. They're simple to understand, and there's no cleverness in the visualization that might hide critical information.
To answer your questions, at Periscope I've seen:
1. A line graph of amplitude over time. You should see the frequency emerge clear as day. If you want to calculate frequency explicitly, you could overlay a second line with its own axis. Again, super simple, but gives you the answer directly.
2. I've seen a lot of fancy graph visualizations, but nothing that makes me happy. Depending on what you want to know about your graph, maybe a simple table with a structure like:
[node name][node name][weight]
[timestamp][node name][node name][weight]
3. OK, obviously I think Periscope is a great choice here. Loads of data analysts use it to visualize time series data on many tens/hundreds of billions of data points.
That said, other good choices are: Excel, R/Stata/Matlab, gnuplot, Apache Pig. And for the data storage itself, IMO Amazon Redshift is unparalleled.
There needs to be a route from the internet to your DB, yes. Sometimes this works because the DB is in the cloud, or sometimes a port or SSH tunnel is opened into a local network.
We plan to treat the email addresses as people who might be interested in our product, and at some point we'll reach out to everyone who signed up and ask if they'd like to try it out.
If you know you don't want to try Periscope, but would like to read the eBook, by all means use the link above. ;)
In the long run, it is a data tool that will eliminate the need for data warehouses.
It turns out the hard problem is not running the queries fast, but keeping the cache accurate and up-to-date. Our cache coherence service is probably the piece of code we're most proud of.
And, yes, we're hiring in SF. :) Reach out to me, harry@periscope.io
BTW, this product (Scalyr) is a lifesaver. We (Periscope) are able to operate ~ a dozen heterogeneous servers with no FT DevOps largely because of Scalyr.
See e.g. https://www.svb.com/private-bank/founder-liquidity/ and http://www.founderscircle.com/services/
This is one of those areas where "being a good person" is a competitive advantage and drives lots of retention. Plus, after a couple months of this, they'll all upgrade anyway as paring down gets harder and more time-consuming.
We went by this: