Performance is really not Java's issue. Even bad Java code is still substantially faster than the bulk of modern software that is based on technologies like Python or JavaScript/Node.js.
This might also be why I heard colleagues saying “Nono, listen, these ‘N+1 problems’ and our nested service calls aren’t an issue because it works well enough” until it eventually didn’t. I’d rather not have bad code in any language.
They are not trying to buy developer goodwill, they are trying to catch up with Antrophic in terms of getting those B2B contracts, which is currently the most realistic path towards not running out of money.
1. The Register reports OpenAI is well ahead of Anthropic in B2B contracts. It's Anthropic playing catch-up, not OpenAI.
2. In any case, the announcement strongly suggests that customer acquisition had little to do with this. The stated purpose of the acquisition, as I read it, is an acquisition (plus acquihire?) to bolster their Codex product.
3. But if they were hoping for some developer goodwill as a secondary effect... well, see my note above.
Measurable responses to the environment lag, Moore's law has been slowing down (e: and demand has been speeding up, a lot).
From just a sustainability point, I really hope that the parent post's quote is true, because otherwise I've personally seen LLMs used over and over to complete the same task that it could have been used for once to generate a script, and I'd really like to be able to still afford to own my own hardware at home.
I'm using local models on a 6 year old AMD GPU that would have felt like a technology indistinguishable from magic 10 years ago. I ask it for crc32 in C and it gives me an answer. I ask it to play a game with me. It does. If I'm an isolated human this is like a magic talking box. But it's not magic. It doesn't use more energy than playing a video game either.
Thanks! I've been playing with some of the qwen models via openrouter as well.. I'll have to give 9b a go at some point, I've been mostly playing with 27b and coder-next up till now.
> Very pleased to see such performance improvements in the era of Electron shit and general contempt for users' computers.
Luckily, we have invented a completely new nightmare in the form of trying to graft machine-usable interfaces on top of AI models that were specifically designed to be used by humans.
I don't know, occasionally there are some funny results. For instance, I have managed to get AWS' support bot to start smack talking their platform and criticising its often needlessly complex and sometimes incoherent design before cheekily offering to help me make my relative simple setup even more complex and enterprise-ready.
They are solid and cheaper, but they don't offer the same level of control plane and API access as Hetzner that is really helpful when managing a larger number of servers.
I am very curios if OpenAI's IPO attempt this year will turn into WeWork 2.0 where all the air suddenly comes out of the valuation once the market acknowledges that they have no moat and lack a clear path to profitability that would make these huge investments worthwhile.
There’s a reason OpenAI and Anthropic are both trying to accelerate their IPO while still being wildly unprofitable. There is still unlimited AI hype in the market. If they go public this year the entire world is going to blindly buy them without looking at their books.
It depends where I look. Among colleagues and tech-native friends, I feel like there’s healthy skepticism as well as the excitement about new tech. On the other hand, all the investment podcasts that I’ve been following for years are nothing but ignorant AI hype and reciting articles about how all the jobs are about to disappear. I guess the people who doesn’t make firsthand experiences are not leaving the hype yet.
Both groups will operate on a wide spectrum, but if we're already generalizing...
Perhaps there's a matter of competing priorities?
Programmers are usually quite cynical overall, but in this case I see it as a "My CEO is telling me _out loud_ that they want to replace me, so why would I help them speed up that process?"
Investors likely want what they're invested in to appreciate, so I imagine they're likely over-leveraged and are doing what they can to get their bag.
> sky-high expectations of AI have come down quite a lot.
they hype for AGI has certainly deflated, i haven't heard anything about that being right around the corner and the implications in a while now. The hype and doom now seems to be coming from software devs only, the front page news articles about AGI have pretty much stopped for me.
/"front page news" to me is the google news, US, Business, and Technology tabs
All these predictions already started aging like milk, for example "Former Google CEO predicts AI will replace most programmers in a year" in April 2025.
The hall of fame will explode with failed predictions this year.
All of these things are vastly overvalued. Only one with tangible value is SpaceX because that's actually a moat-space. OAI holds no moat, has not done a good enough job to entrap their users, and has poor cost structure.
xAi isn't even a point of discussion.. it's just a scheme to rip off investors.
WeWork.. hard to take anyone seriously that ever invested in this bad boy.
Wework was a valid long bet that office properties would re-appreciate once the pandemic stopped — but now that AI is pulverizing the job market, any hope of that long bet paying off will require one of three things: a free-market boom in workers that require commercial property for success (e.g. physical invention companies like e.g. Saildrone due to not being able to homelab resins for safety reasons), and/or a market-wide rehiring event due to AI’s failure to deliver, and/or regulatory shifts in profit taxation and new business investment that trigger the above-described boom.
I know some commercial property owners in my hometown let their lowest-desirability storefronts sit vacant for twenty or thirty years (!) in order to prevent commercial property rent from falling across their entire portfolio. Turns out you can pay a lot of property taxes with not much revenue, and there hasn’t historically been regulatory pressure to pay an escalating “empty tax” to compel landlord pricing to behave according to supply and demand pricing models. Wework is still a terrible investment for an investor, but if you’re looking to bet long with no call and have the patient of decades, it’s not the worst plan. (There are certainly worse ways to gamble your money on the commercial property market!)
> any hope of that long bet paying off will require one of three things: a free-market boom in workers that require commercial property for success (e.g. physical invention companies like e.g. Saildrone due to not being able to homelab resins for safety reasons)
That doesn't make sense for WeWork, though. Aren't they a rent-a-generic desk company? If you have any kind of specialist requirements (e.g. "processing resins") they'd seem like a bad fit.
SpaceX valuation is also going to be interesting. Talking about CapEx, SpaceX has deorbiting assets on top of depreciating ones. And without Starlink the space launch market size is pretty small.
> SpaceX has deorbiting assets on top of depreciating ones
The deorbiting part is redundant. Their satellite are just that, a depreciating asset. Their lifetime seem to be 5 to 7 years. The important claim is if the total cost, including the launch, can be recuperate over that lifetime or not.
Is it? It’s less than obvious that the orbital datacenter boom will happen. Space mining could be a big deal but that’s not a foregone conclusion. Maybe someone will want to build a huge radio array on the far side of the moon, but I don’t expect hundreds of billions to be spent launching it — who would pay? Mars is less fashionable than it was a few years ago. Starlink is pretty impressive, but so is boring singlemode fiber, and the latter is increasingly being deployed everywhere.
(Obviously it will “skyrocket” unless someone comes up with a commercially viable launch system that doesn’t involve rocketing into the sky…)
There is Starlink and Amazon’s version, which both need many thousands of launches each (ignoring data centres)
There’s US military stuff like the starsheild, and it seems extremely likely china will follow suit.
Sat internet is changing the world rapidly where fibre can not be run. I saw it first hand in remote Australia, Yukon, Alaska and Africa. Not to mention ships and planes.
These projects are aggressively driving down launch costs, which increases demand for launches, which drives down launch costs which drives up demand.
Yep, SpaceX actually has a track record as an actual leader and innovator in it's niche (that's very CapEx intensive to enter), it's not really a moat but it's a lead that no other entity seems to be closing in on (on the contrary many would-be competitors seems to have almost given up).
As for OpenAI, I'm not sure if Altman is an idiot or fraudster, claims about reaching AGI/ASI with scaling and investing in that fashion was always delusional at best or fraudulent at worst, maybe he just hoped to divert enough money to engineers to make actual breakthroughs or that the hardware would become a moat but competitors have kept pace, and I fully agree that they are mostly now only hanging on with an insanely bad cost structure now.
> on the contrary many would-be competitors seems to have almost given up
maybe the smaller ones; Blue Origin succeeded, and French and Chinese nu-space companies will continue to get funding for decades - national governments are capable of footing the bill of large CapEx projects. SpaceX competition is irreversibly tied to US foreign policy, and only scientific amd commercial launches are price-sensitive
Twitter is already dead as everyone on Hacker news knew. Nobody I know uses X or whatever it's called now, xAI? I'm looking at Musk going bankrupt and as soon as that happens Trump will be Impeached.
They would need to have massive accounting fraud and lose public support. Unlike Enron, Nvidia actually sells tangible goods at a massive profit, and hasn't appeared to gloat over people getting screwed over.
> They would need to have massive accounting fraud and lose public support. Unlike Enron, Nvidia actually sells tangible goods at a massive profit, and hasn't appeared to gloat over people getting screwed over.
But what happens if they can no longer sell those tangible goods at a massive profit, and they gave return to their roots selling to gamers? When the boom ends is when massive accounting fraud could happen.
i think worst case they go back to selling GPUs to gamers. That would indeed be a massive front page story and movies would be made about it but I don't think the fraud is there. They have a pretty straightforward business, make GPU cards and sell them.
"They have a pretty straightforward business, make GPU cards and sell them."
They do, but that's not the (full) story here. Companies tend to easily migrate upwards, to a higher volume and/or higher profit margin market, and hardly (if ever) in the opposite direction. The painful restructuring necessary to enable this kind of reverse change is also damaging to the company brand, culture, and self-perception. If they ever get in such position, they may of course recover, but I wouldn't bet money on that.
> make laws that prevent a duopoly, penalize anti-competitive behavior and push open-source standards for software/hardware.
None of this is legally easy to implement or enforce, and any attempt of doing it is virtually guaranteed to create an unbelievable amount of unintended consequences as people figure out ways to game this new set of rules.
We need something similar to FIPS for interoperable software and standards. Organizations will fall in line when money is at stake.
Say for example your local/state/federal agency publishes (or accepts) documents exclusively in ods/odf instead of proprietary formats, that will automatically drive adoption of software and prevent lock-in.
Agressive interoperability at the protpcol and exchange format - its why email mostly works even forcing Google to back off when they tried to change email to be rendered by their cdn (i forget the name of the offering - but was similar to what news pages were being pushed for speedup).
Bad actors will always abound - like Microsoft spiking the documnt standards by pushing through ooxml when odt/odf was gaining traction.
Or basically just coercing the decision makers like in Berlin(?) where they moved their offices into hte city to get them to drop Linux/Openoffice.
As i rmeber it ooxml backers made it intentionally harder to parse the specs than was necessary ,if it was fully open i believe the open source implementation would have been on par. As it is its subtly broken in annoying ways , and with Word being the default - its version wins out and gets to be the only acceptable submission format.
If you notice most doc submissions when its not a pdf being requested will specifify MS's version.And by sheer momentum the alts get less traction.
I think their main issue is that they seem to have no one who is seriously looking at the Matrix ecosystem from a product perspective. You have all of these pieces of technology in various states of maturity that more or less fit together if you know what you are doing. But there is also a lot of friction and a lot of things breaking on a regular basis etc.
What the project needs is someone who looks at it from a customer perspective and who can direct resources to make sure the entire thing is packaged as one consistent thing that does what the customer needs.
If you install WA or Signal, or if you sign up to Slack, you don't have to wonder which home server you should install and which of a dozen or so available clients you should use and what features are not yet production ready. Instead, it just works.
The lack of attention that you identify is a real issue with the project but the root issue is ultimately a lack of sufficient funding that would enable all these parts to receive the attention that they require.
Funding fixes all these problems and it has to come from big governmental and institutional players in Europe who are motivated by ending their reliance on American companies like Microsoft.
I think a product focus does exist: Element seems to be a genuine attempt to fully assemble Matrix as one full project. The problem is that it feels like the Element devs are stuck wanting to have their cake and eat it too.
There's some design choices in Matrix that don't really "fit" with what modern messaging infrastructure looks like. (Which to summarize it pretty quickly is a Slack/Discord-esque model, where non-sysadmin users get to fully administer their own spaces, with an expectation for multiple different channels, control over user permissions and user access and so on and so forth.)
Some of these come from the fact that Matrix is pretty blatantly just designed as "what if IRC, but slightly more modern". It's main unit for non-sysadmin moderation is a single channel, with the expectation that one instance of Matrix will never have two channels named #general (as an example). Similarly, it's entirely possible to kick users from a channel... but then have that exact same channel continue independently on a different instance, but under a different label. This makes sense if you look at it as "supercharged IRC", but becomes a complete and utter mess when you factor in things like the encryption between two servers suddenly disagreeing with each other (leading to a bunch of old messages becoming unreadable), content moderation (barely an issue on IRC because message retention is expected to be almost entirely clientside) and so on and so forth.
Element/synapse's people do try to provide for these cases, but you're effectively stuck trying to prod at admin API endpoints, bots to synchronize moderation decisions and they have like 3 different "channel grouping" that's supposed to be their version of the Slack workspace/Discord guild model.
Honestly though, I'm pretty sure that once XMPP gets a proper multi-user multi-channel XEP going (there's one in draft right now which specifically tries to provide workspace-esque support; it's possible to do this already but it's a sysadmin XEP, the proposal aims to give this capability to regular users), it'll just end up blowing Matrix out of the water entirely for most usecases. Unlike Matrix, it's a far more mature protocol that's a lot easier to work with and actually has many different implementations that you can choose from.
Room addresses/aliases (like #matrix:matrix.org) must point to a single room (in fact, they point to a room version, so when rooms are upgraded, addresses must be pointed towards the new room). But for communities, a better way to organize the rooms would be spaces. Spaces can be joined. Spaces can contain rooms and other spaces. Like discord "servers" (guilds), but more flexible.
reply