Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
OpenAI Launches Smarter, More Conversational ChatGPT 5.1
- Thread starter MacRumors
- Start date
- Sort by reaction score
So did it actually get better this time, or is the OpenAI press release just media spin, and in reality they just made it use less tokens again?
I disagree strongly, and unfortunately your knowledge regarding Anthropic isn't exactly correct on this: they are working with fluidstack and are invested in Google TPUs, it's got nothing to do with NVDA at all. Amazon is their financial backer much like Microsoft is for OpenAI, and both aren't going to give up within the next few years.
People really want this to be true and are hoping for some crash soon for a lot of reasons that I understand very well. I also understand the technology pretty well, and have spent a lot of time working with research technology precursors. There are limitations but the utility is undeniable, the usage is upwards of 1 billion people, and the US DoD has large contracts with all of the foundation model companies.
GenAI isn't going anywhere, whether we like it or not. And there are a lot of reasons to not like it, but it's the reality.
Ads are coming and the infrastructure is already there. For coding etc. the prices will go up (Anthropic's prices are already high and profitable for that) and companies will pay it because the utility is there.
Free users will be milked for data and ad sales. Paid users at lower tiers will probably also be ad targeted too. It's obvious what's coming near-term.
Long term we've got World Models and other things to look forward to. Yann just left Meta to start his own company for a reason
.As far as "dumping onto bag holders" nvidia is somewhat risky but the ROI is enormous, they have $600B in commitments through 2026 which is ridiculous. Of course that won't continue indefinitely but their technology and particularly their ability to build out fast with clusters and their specialized networking hardware / software stack is unmatched.
The risk comes in now that China is looking to ban imports of the chips, because eventually there may be a competitor or some novel technology that emerges from there which takes the lead away. That is a risk for investors, for sure.
All of this is a risk too, as I said there will be a crash absolutely once the small companies fail and people misunderstand the long-term... but the smart money will hold on (or dump high, and reinvest low, which is some of what's happening now).
Feel free to @ me in a year when NVDA is in the toilet, I'll take my naive self and eat some crow then!
I am blown away by the sheer volume of negativity in this thread, so obviously I haven’t been paying attention 🤷♂️
So many speculators here!
I use ChatGPT for loads of things, including coding in C++, and it’s very helpful. Not always 100% accurate but helpful. I suspect many of the folks here ask ChatGPT to create miracles of code out of 10,000 lines of garbage and are frustrated when it doesn’t work. In my experience, giving it simple, atomic chunks to work with is great. It understands, it helps optimize, and it often gives multiple solutions. But if you, the user, don’t know how to code, really, not even ChatGPT can help.
Or it could just be that you’re using python.
So many speculators here!
I use ChatGPT for loads of things, including coding in C++, and it’s very helpful. Not always 100% accurate but helpful. I suspect many of the folks here ask ChatGPT to create miracles of code out of 10,000 lines of garbage and are frustrated when it doesn’t work. In my experience, giving it simple, atomic chunks to work with is great. It understands, it helps optimize, and it often gives multiple solutions. But if you, the user, don’t know how to code, really, not even ChatGPT can help.
Or it could just be that you’re using python.
If Sam Altman is involved, no thanks. OpenAI is as shadowy as Alphabet, Google, whatever they’re calling themselves these days. Grok is the superior platform, anyhow.
I was with you until “Grok is the superior platform, anyhow.” It’s hypocritical to shame AI built by Altman and then praise AI built by Musk. Have you seen Twitter these days? Or how Musk is trying to “fix” the opinions of Grok? Talk about shadowy... I believe they both deserve the shame.
I don’t use Twitter, so not sure. Elon isn’t much better, but he doesn’t hide his intentions and is pretty straight forwards about what he has in mind. Sam Altman on the other hand. I wouldn’t dare turn my back on him if we were alone.
You have a point there. Elon is definitely very transparent about his goals lol. I didn’t see it that way. Thanks for pointing it out.
Have not used Grok much but I like Gemini as it is hard to realize it is machine (on one side scary) but.... GPT provides sometime more precise and useful information but in very rough form. That is my experience. So I use them in tandem.
It is natural evolution of growing computational power and one of areas that drives sales of new HW otherwise for majority current performance is for majority overkill.
But since many thinks technological advancement = to progress of society here we are.
If they will not find proper ways to monetize all investments other that state surveillance and military purposes then blast will be audible across Earth lol
All good points, but not just for LLMs. This is true for any technology we use. Adobe Creative Cloud becoming crapware. Affinity going into the Canva Lockware-Ecosphere. Office the same. Cloud spaces like Dropbox bring central to your work yet obviously rudderless. Operating Systems ****ing up SMB. Apple becoming worse every year, Google however being much worse already. Your internet provider has problems and Mail, Downloading and anything online will not work for hours. Working with tech makeshift work more fragile and dependent. Which is why security driven IT seems to bring the companies they work at back to stone-age level of tech in order to reduce risks that cannot be avoided as they are tech-inherent. The only ones safe from all of the above and more problems of technology are people working without, craftsmen and artisans, small shops, artists, writers and so on. The less tech you use, the safer. If you work in or with. A lot of tech or software, welcome to the worries where every update can mess up your business.
Still, AI is neither a bubble nor only a assembly-line-style method of replacing human labour force. Iris neither a glorified spell-checker nor a doomsday machine. The fact that it is allow the above and much more, the speed of implementation, the global competitiveness and the fact that we live in times of anti-democratic uncertainty generally make the technology a very unsafe ground to operate on and En********ation is absolutely sure to happen sooner rather than later.
But that doesn’t change the fact that this could be the most game-changing technological innovation of our generation and we all should not be Luddites but embrace it, shape it and as consumers, designers, users, devs, managers do our best to make it as good and as beneficial as we can.
You raised a great point yourself that made me think! I appreciate it!
"More conversational" = The techbros can't live without some virtual sycophant telling them they're right all the time.
that may well be, but the AI genie is not going back into the bottle
The previous AI winters were almost entirely B2B/institutional, which is an incredibly different environment from the AI market of today.
The second fundamental difference from then to today is that current AI has crossed utility and mass market adoption thresholds that makes it sticky in ways previous generations didn't.
A "winter" will undoubtedly happen in the investment/hype cycle of AI, but the consumer adoption creates a floor that prevents complete collapse. It's more like social media - even if investor enthusiasm waned, the products were too embedded in daily life to disappear.
AI is a foundational technology. It will be around much longer than you or I, regardless of what the markets think about it.
There is virtually no consumer adoption.
The moment you ask money for it, the customers go away.
It doesn't solve any useful problems for the average person.
It's dead.
The mathematics and the idea will live on, but the commercials mean it is dead.
A BILLION people are using GenAI.
Facebook doesn’t ask for money and LLMs specifically are going to be monetized similarly for product references and advertisement.
OpenAI introduced the product carrousel for exactly this reason and are already providing tracking links ahead of the affiliate programs.
This is just patently untrue, if you have data share it, instead of what you might understandably want to be true.
Reading doomsaying takes without a critical lens or understand the technology or how it can be useful is judging something from a place of ignorance, not a strong foundation of understanding.
Consider how many people even knew what “AI Winter” was in the late 90s, or expert systems before that, etc.
The scale, market penetration, and technology is wholly different this time.
“AI” is still the wrong name for the nerds who understand they really don’t mean “ASI” but for average people, all of this is AI now, right or wrong.
LLMs are a tiny part of the potential for AI. In addtion to Small Lanuage Models and Private Small Language Models, many forms of AI being developed today are not langage models at all. If you don't believe me, ask your LLM.
LLM today feels like when the EU was consumed by the fear of Microsoft bundling Internet Explorer with Windows. They thought the browser leader would dominate the internet.
What they didn't see was the coming revolution. For example, bank payments processing (check and wire transfer) as well as policy sales and claims processing at insurance companies. It transformed our customer experience, greatly reduced costs while absolutely gutting hundreds of thousands back office of jobs in the US financial industry, plus at all companies that billed by mail and an estimated 300,000 at the USPS. Who won the browser war had no impact on that.
If I’m in a noisy environment I will mute the microphone as soon as I’ve finished speaking so the background noise doesn’t cause chatGPT to stop and wait for more input.