> And I look down at my arm and I realize that I can no longer define the boundaries of my body. I can't define where I begin and where I end, because the atoms and the molecules of my arm blended with the atoms and molecules of the wall.
Speaking of backup solutions, why don't we ever see pro-sumer priced tape drives? The technology behind capable LTO drives is now more than 10 years old, shouldn't we see some reductions in price now?
> why don't we ever see pro-sumer priced tape drives? The technology behind capable LTO drives is now more than 10 years old, shouldn't we see some reductions in price now
First, LTO drives are conceptually simple, but if you've ever opened one up you'll know they are a feat of horribly complex engineering.
Second, they are not a commodity product. Infact thanks to the magic cloud they are even LESS of a commodity product than they were 10 years ago because lots of people have either wholesale moved to the cloud or use S3 for backup.
1 + 2 = Low volume product with lots of parts crammed into it = high manufacturing cost = high price.
The orgs who still use LTO in their infra are the sort of orgs who don't blink at the price tag. The cost of the CTO's farts is probably more than a 5k tape drive.
Back in the era when Zip drives were around, there was a prosumer/SMB tape standard: 1/4-inch QIC. I sold many PCs with a tape drive in one of the drive bays. They were awesome.
They also sucked from a support point of view. People would mistreat the tape. Store them next to giant magnets or on top of the microwave in the restaurant. Forget to run the weekly backup and then blame you when they lost files. I couldn't wait to get them on to CD-RW which showed up for SMBs very soon afterwards in the late 1990s, and then eventually to the cloud. What a relief to no longer need magnetic media.
The irony now is that many of the SMBs I see today (though I no longer consult for them) have effectively zero backup because all their business process is tied up in a SaaS that they do not control. Eg their website is on squarespace, their tasks are on Asana, and their finances are in Quickbooks. Any one of these goes dark, or out of business, or is vandalized and it's curtains for the whole business.
My experience with tape is very out of date but I doubt much has changed due to the nature of tape
Tape software sucks. Tape restores are cumbersome. A SMB can literally but a multi TB HDD and just drag and drop, and but the drive in a safe.
A SMB will need to hire someone who wants to deal with this niche tape storage, and why would they bother if they can use a NAS for that and a dozen other things at the same time
Tapes can have a shelf life of 30 years. If you need archival storage then tape is a great solution. For everybody else it’s more trouble than it’s worth.
How many of them have actual need for terabytes of storage? And I mean business continuity critical one.
Overall they are much better of buying a few external HDDs. Standard interface, sufficient capacity. Just get trusted person to carry one out every month or three months. For continuous stuff just have NAS.
> millions of small-to-medium businesses that could use a good and cheap backup solution for terabytes of data.
Yes, and that's exactly how small-to-medium business IT used to operate.
These days sadly most small-to-medium businesses are drinking the cloud koolaid.
You would be hard pushed to find a small-biz with a comms room these days, when back in the day every half-decent small-biz office would have a comms room with cab and a few servers in it.
Now most small-biz are on Microsoft or Gmail for mail and their office is full of laptops on WiFi.... nobody has any respect for good old-fashioned structured cabling these days either, sadly. ;(
It is what it is, sadly.
Hence only governments, enterprises and, ironically (if rumours are correct) the cloud providers (for their archival S3 products) are still buying tape.
Unfortunately, I never got around to using it! I bought it with high hopes, but my Zip disks turned out to be so convenient and spacious that the VHS-backup need never arose.
That said, it's not too late... I still have my Amiga system in storage, and a VHS recorder.
>I bought it with high hopes, but my Zip disks turned out to be so convenient and spacious that the VHS-backup need never arose.
It's good to hear, in retrospect, that you were able to use a storage medium that did not even exist when Amiga were discontinued. Which type of interface for the Zip drive works with it?
(It occurs to me that Zip disks presumably offer the great virtue, otherwise absent as I understand it for Amigans, of PC compatibility.)
The reason we open our client side code is to bring in the trust in putting rtrvr's DOM intelligence in your web apps - https://github.com/rtrvr-ai/rover/tree/main . Our monetization is super straight forward with subscription - https://www.rtrvr.ai/pricing . The experiences of some extensions shipping anything or selling user data comes in when people build them as side-gigs not when we pour more than year in building the highly accurate automation engine. We have cloud sandboxes too if you prefer executing with the same intelligence on cloud and not on your own device.
auditing the code is fairly straightforward if it isn't obfuscated. so long as it doesn't execute dynamic code that is. but the big issue is you can't control when the extension itself gets an update (to my knowledge). and it isn't uncommon to sell browsing data, or the extension itself to someone more shady than the original author down the road.
Insurance works on repeatable, predictable risks based on models.
You can't get insurance against a military attack. But you can hedge using a prediction market. It's essentially the version of insurance for one-off events, that relies on wagers since you can't use models.
I'm not saying it's not out there, but in many cases finding insurance for act of war is difficult. It is certainly rare and almost always far less accessible than using polymarket. If you have a very large holding you can probably get someone to write it for you though.
Of course they will. Tokens are valuable, you can always spend a finite budget on specialized tokens or fewer and higher quality tokens, size of user base and engagement gives you a flywheel moat that is difficult for newcomers to compete with. The market is complex and easy to oversimplify.
My new startup tokencoin will blah blah blah exchange rate, (something AI writes here), 3. profit (more AI), benefiting all human kind and helping our users scale up their productive intelligence!
It's hard and complex to enter any mature market. The vast majority of firms that attempt to enter a new market fail. LLM's have no more than this normal moat.
If every market has a moat, then saying that a particular market has a moat is a statement without meaning. The OP was probably not trying to make a meaningless statement, therefore the OP was probably saying that LLM's don't have an abnormally effective moat.
I agree, LLM's don't have an abnormally effective moat, just the standard moat most mature markets have due to market complexity. IOW, LLM's will likely end up with the standard oligopoly most modern western markets end up in, which have minor but relatively ineffective pricing power.
Isn’t capital and momentum a moat? Sure Chinese models use distillation but I don’t see them training models from scratch. At least not today. But maybe as chips get cheaper and they have Chinese made ones?
Apparently not much of one. There are, what, 5 or more companies with frontier models? And open weights models like MiniMax are snapping at their heels
There are many markets where open source has been nipping at heels for a long time.
Obviously product areas differ for reasons structural and happenstance. But there is definitely a pattern that occurs, where open source fast follows commercial advances, benefiting from having a clear target to develop for.
Which is of course, a great service. Even if it never unseats the commercial version, it forces the owners to reinvest more in improvements, by undermining their moats. As well as providing a much better value alternative version for many people.
And it does not even consider that e.g. the EU might one day decide that AI should be for everyone, thus releasing a heavily subsidized open source model.
Or that at some point AI is good enough, and so at that point any model will do.
I’m not technically familiar but I remember someone saying that models like MiniMax basically skip the cost of training by using distillation to basically “steal” the models from OpenAI or Anthropic, and that these companies now have various defenses against this. What happens when MiniMax has to do the full work themselves?
On top of that, as long as big companies don't take the protection of my personal information seriously, why should I worry about violations of copyright laws? It works both ways.
I'm betting on a company like Taalas making a model that is perhaps less capable but 100x as fast, where you could have dozens of agents looking at your problem from all different angles simultaneously, and so still have better results and faster.
Yeah, it's a search problem. When verification is cheap, reducing success rate in exchange for massively reducing cost and runtime is the right approach.
I'm excited for Taalas, but the worry with that suggestion is that it would blow out energy per net unit of work, which kills a lot of Taalas' buzz. Still, it's inevitable if you make something an order of magnitude faster, folk will just come along and feed it an order of magnitude more work. I hope the middleground with Taalas is a cottage industry of LLM hosts with a small-mid sized budget hosting last gen models for quite cheap. Although if they're packed to max utilisation with all the new workloads they enable, latency might not be much better than what we already have today
reply