LLMs are great at aggregating docs, blogs and other sources out there into a single interface and there has been nothing like it before.
When it comes to coding however, the place where you really need help is the place where you get stuck and that for most people would be the intersection of domain and tech. LLMs need a LOT of baby sitting to be somewhat useful here. If I have to prompt a LLM for hours just to get the correct code, why would I even use it when the tangible output is just carefully thought out few 100 lines of code!
> A lot of the big ad networks right now instead rely heavily on geo-data
How does this work in today's age where ISPs normally will have at least one level of NATing with ipv4. And given ipv6 with prefix delegation is still far away this should continue to be very imprecise?
Not sure about US, but Indian ISPs are doing this already to conserve IP space given huge userbase. In theory it would work similar to how a NAT gateway works for outbound communication. Skan + geo would be hard nut to crack in India.
In UK I'm now on FTTP but even on ADSL the house would have an IP address that normally stayed constant until a router reboot. This seems to be pretty common in the UK. Probably on cable internet (and on mobile ofc) you get NAT-ed but I've never had that.
It still works because those CGNAT shared IPs still vaguely correspond to a certain geography. It won't be accurate enough to target a specific home, but still accurate enough to target a specific neighborhood, for instance.
Assuming an ext-IP (60k ports) can easily represent 100 household if we statically assign ports. Given CGNAT with dynamic port allocation this can easily go up to 5x? That's wildly inaccurate given the core problem is to "target" a small set of users which is based on this geo info. Not sure how well this elephant sits in a room full of engineers solving this specific targeting problem.
Billboards are still among the most effective forms of advertising in terms of efficiency. You don’t need to be very close. I see myself popping up probably 10 miles from where I’m actually at, but the businesses aren’t that inaccessible.
Reminds of the old days, when reverse engineering game binaries were a thing. Finding hex strings, no-op'ing if conditions, adding jmp instructions. Many heroes forgotten with time as more and more of software onboarded the cloud offerings. I wonder how much of it is still relevant and in what fields?
It isn't, but it's not that far from being a serious comment.
The old man has an ancient PC choking on dust under his desk that runs a QWS3270 terminal that updates the mainframe directly. If it's a big bank, the bank owns the mainframe. If it's a small bank, a vendor (probably Fiserv) does.
I misread this for some reason and I am now interested to know if DNS can be made to work dynamically this way instead of being just preconfigured static records?
This is excellent, kudos to the author. I wonder if pg exposes enough control to develop a tool along with this information which help analyze locks and their effect on queries on a live server.
Moderna vaccine also seems to have similar storage requirements with a high price tag. BioNTech otoh has impractical storage requirements for the third world. AstraZenca's price and storage combiation may open possibilities for distributing this globally.
Encrypted SNI has been in talks for a long time now. It was meant to be out now but it was shot down at the last minute. And yes, its the last thing that allows a MitM to work out which site you are on if its hosted on a shared hosting platform. Likely this is the reason its hard to get approved.
When it comes to coding however, the place where you really need help is the place where you get stuck and that for most people would be the intersection of domain and tech. LLMs need a LOT of baby sitting to be somewhat useful here. If I have to prompt a LLM for hours just to get the correct code, why would I even use it when the tangible output is just carefully thought out few 100 lines of code!