The product itself is very good, but the sales process is truly awful. Random calls with non-technical reps unable to answer basic questions like, "now that you've added this to my GCP account for 2 weeks, how much is this going to cost?" They'd say they're not sure but they have a startup deal with $xxxx minimum commit for 12 months gets you two months of extra trial, cancel anytime no questions asked. It's not just bad, it's comically bad.
I think it just depends on what you're working on. In my experience, Claude usually has the edge over ChatGPT for most coding or logic problems; however, there are situations where Claude goes too deep the wrong way and ChatGPT wins out by staying more high level. Claude and ChatGPT are also "free" too, btw.
I've had mixed success with evading regional locks for sports and other content with random VPN providers. Any cloud provider I've tried is immediately blocked. Pretty easy to do since they all publish their IP ranges. If you have friends around the world, an easy thing you can do is have a shared Tailscale network with exit nodes.
Running a VPN out of cloud services pretty much guarantees recaptcha prompts at most sites and denied access to most streaming platforms. Plus, paying per-GB adds up fast. If you're trying to evade region restriction or sail the high seas, stay away from cloud providers.
This has been my experience with any model with a large response token limit. I've had to work around this by running it through several times with specific questions about the data: extract text, extract tables, extract <specific detail>. They seem to do well on large input though so I just concat all the extracted info and things seem to work just fine.
Good for them! I hope this gives them a bit more time to R&D, stiffen the chassis, etc. before taking it to market. But, I also hope Formlabs doesn't raise the price as an entry option to their 5-figure Fuse 1 printer. If they can keep the target price with a Formlabs logo, it'll be a hit.
Can you? I tried to do this, even asking their AI support and it replied with something like, "Substack was built to help writers build their audience. If you are not looking to build your audience, Substack is not the right product for you." So lame.
This is pretty cool. For folks on GCP, does it make more sense to use vanilla rails with cloud run? I wish lambda had the ability to share processes like cloud run with a max concurrency setting.
I recently saw a post about Cackle [1] coming out of the rust ecosystem which looks pretty cool. Similar to what others have said, there's always a chance that a C lib or shell script is leveraged to bypass compile time guarantees. I think that's why it's important to do both. Do what you can to keep your supply chain safe and limit your runtime as much as possible.
The ability to run a cron every few seconds is pretty cool and avoids those OS cron scripts that loop and sleep, so I'm glad that was added. Only downside to cron jobs running inside postgres is PL/pgSQL. It's never my first choice for scripting sql operations. I'd much rather use python, ruby, etc. I'm glad it exists when there's no other option, but it's kind of its own beast.
That said, I'm very happy to see more and more stuff coming from Citus. We're looking to split a few growing databases and a bunch of folks are excited to try Citus again now that the enterprise features have been open sourced.
Vertically scaling has a cost component. If you're able to use pgbouncer with transaction pooling to better utilize but not exceed your existing vertical postgres limit, then you're set. If you do, then you must scale up. Pgbouncer can help you lengthen your runway. Tuning your pgbouncer pool sizes can help minimize the overhead but as with anything we do, it's about weighing trade-offs for your situation.
This. For the web service workloads most of us run, I’ve always run out of postgres connections before exceeding the vertical limit of a postgres server.
The real problem is working with postgres’ “each connection is a process” model. Pgbouncer puts an evented pool in front of postgres to deal with this. Apps that are aggressive with DB will not benefit from having an evented pool in front. However, web apps (think rails) will have connections checked out even if they don’t need them. Pgbouncer helps here. If your app recycles DB connections when not in use, that leads to connection thrashing and higher latency, which pgbouncer can help with. But you’re right that at some point, the DB is the bottleneck. For most people, it’s the number of connections, because postgres makes a process for each connection.
Daily driver for me as well, but only for terminal. People laugh sometimes while pair programming, but usually by the end they begin to really like it. Can't use anything else at this point.
Corruption of database storage is the biggest fear so most upgrades are done cautiously. Historically postgres versions were released much much slower so the upgrade cadence was slower too. That said, we've been upgrading more frequently, especially since a lot of performance features are being released. So far our upgrades have been executed flawlessly for our cluster of postgres servers (~30 postgres chains each made up of 4x large dedicated servers).
We tried for multiple years to make Nomad work because it's simple. We're already enterprise customers of Hashicorp, too. We love Hashicorp! Nomad is a great scheduler and the UI is wonderful, but there is no story around network security. Example: You have postgres running on dedicated metal, and you want to restrict which nomad services have access to postgres. Consul Connect allows you to enforce access policies, but these work more like haproxy with automated service discovery. There is no batteries-included way to prevent traffic from leaving your container (or identify which container the traffic is originating from). You can use a custom CNI plugin to give each container an IP address from a unique subnet per service (and restrict traffic at your firewall/ selectively add subnets to iptabels on your postgres server), but now we're adding bespoke complexity. We brought this problem up to the Hashicorp internal teams over many calls, but ultimately they said we'd need to pay for a 3rd party overseas consultant to advise us on implementation. They weren't sure what to do. K8s is certainly more complex out the gate, but you don't feel like you're the first person in the world to implement.
That said, I think Nomad is a few years away from being something truly amazing. If they can solve the network security problem (beyond optimistic documentation), I think it'll be amazing. For now, it's just a good scheduler.
The Financial Data Exchange (FDX) working group is pushing for this. They have a data spec and UX guides that encourage this exact thing. A user creates an access token for a specific app and that access token might can grant certain abilities to certain bank accounts. It's also adopted by the top US banks (live in production).
I was recently visiting family in southern california and every FBO wanted a minimum 2 hour checkout, and a minimum 1 hour of ground school covering the LAX/SNA airspace. There was an additional checkout process to visit Catalina Island, and the FBOs "strongly discouraged" flying with pax and instructor before the checkout process had been completed. All to fly a C172. Crazy amount of friction. I'm grateful OA tried to solve this problem even though they weren't in my home area.
I've been sponsoring a channel for a few months now but still find that youtube runs ads on their videos. I guess the perks are for exclusive live chats and stuff. Still, it's better than not having a sponsorship system.
> The modern Wikipedia hosts 11–12 times as many pages as it did in 2005, but the WMF is spending 33 times as much on hosting...
Wikipedia also keeps "revision history" and a "talk" section for each page. Those original pages created before 2005 can still be modified, potentially adding to the overall cost. Breaking news stories can cause massive amounts of churn which I imagine has increased with the site's popularity [1]. So really 11x pages = 33x higher hosting costs doesn't seem unrealistic considering how much metadata is associated with each page. That's not to say there isn't a problem, but "page count" might not be the best metric. I wonder what average number of revisions per page looks like over the last 12 years.
Agree 100%! I also find it weird that you can only be signed in to one web client at a time, and your phone MUST be connected to the internet for it to work. Telegram also has a really great developer experience while WhatsApp is locked down. Hopefully biz verification is the first of many new product features. I know making big changes with as many users as they have is really hard, but I would love to see some momentum.
Does github report RCAs for outages to the public? I'm wondering if these outages are related to their migration to kubernetes. Either way I'd love to learn what's going on and how they resolve these incidents. There's a lot to learn and I find these things super valuable.
In the LinusTechTips review he mentioned only top 5% dies go into Threadripper. So if it's related to quality of Ryzen, then that might make sense. I'd be surprised if Threadripper isn't plagued by this bug, though.
I agree. My point wasn't necessarily that they should charge for storage but more so allow the community to support them financially as they're working on their product(s). The post saying, "we're not collecting money, we're not trying to make money right now" just seems kind of strange. Keep the free model. Let people play around, but I still don't see harm in allowing someone from the community to pay keybase for working hard on a hard problem. I'd be more than happy to pay them right now just because I love the work they do and want them to succeed.
This looks really cool. I'm a little concerned that "there is currently no pay model, and we're not trying to make money. We're testing a product right now, and we'd like to bring public keys to the masses." What's the harm in charging $5/month for 50GB as a beta service that might not last forever to help fund development? Maybe even keep the same level but allow people to pledge monthly to support your efforts? It just always seems to be the case that there's a lot of runway until there isn't. Building a product and making money are not mutually exclusive.
MX is not owned by any bank. It's still privately owned startup. We are long term in that we have many 5-7 year contracts with big name banks and add more each quarter. We offer mobile banking solutions, data categorization and budgeting tools to banks, so we Atrium is a product to sell access to the services we built to power our other products. So yes, our product will definitely be on the market for years to come.
We do add as many data sources as possible. We can use Morningstar for example to find cusip of holdings data. We're always trying to improve too :).