Looks like Azure is experiencing a major outage, but I cannot find anything about it.
If you look at downdetector.com you'll notice reported outages from OpenAI, Microsoft 365, XBox Live, Walmart, the list goes on.
>Impact Statement: Starting at 18:44 UTC on 26 Dec 2024, you have been identified as a customer who was impacted by a power incident in South Central US and may experience a degraded experience.
>Current Status: There was a power incident in the South Central US AZ03 which affected multiple services. We have applied mitigation and are actively validating recovery to the impacted services. Further updates will be provided in 60 minutes, or sooner as events warrant.
The times are the same for OpenAI - first notice from 11:00 PST (19:00 UTC)
Power outage seems really odd. Don't datacenters usually have multiple redundant power supplies + on-site backup power generation? Maybe power "incident" mean something else?
The switching equipment can fail. Had this happen at a DC where the switching equipment arc flashed when going from mains to diesel generators. The switching equipment detected the arc and then locked out until someone onsite could inspect the equipment and override it. The rack UPSes only lasted like 5 minutes and then everything went dark.
I had equipment at a colo facility -- that had a (licensed, bonded, not fly-by-night) electrical tech accidentally drop a tool into the main bus connecting mains, generators, and batteries.
They are lucky they were able to walk away, but the facility was dark till someone could get in there and give the power equipment the green light.
Things happen, e.g. the redundant system also failing, the system that should handle the failover failing, a short circuit that causes enough chaos that the redundant supply shuts down rather than feeding power into a potential fault, ...
If nobody minds a plug: My own product, StatusGator, which was launched here on HackerNews 10 year ago, notifies IT teams about outages before they are acknowledged by official status pages.
- This OpenAI outage[1] we notified 4 minutes before they acknowledged.
- The last AWS outage[2], we notified 28 minutes before they acknowledged
- There is def an Azure outage[3] now yet they have still not updated their status page. We notified 35 minutes ago.
Btw, tried to sign up and got a message that it would send me an email to confirm my login. Instead what I received was an email pointing me to a video demo. Not sure if simply clicking on that link was enough to confirm my email. That’s outside of a normal workflow.
Wow, I subscribed to pro-mode and have been using it like crazy since release to assist my firmware development. Extremely helpful. Now it's changed my workflow so much, I don't even want to work when it's down.
Would you be willing to elaborate on the ways in which the $200/month subscription is better than the $20/month subscription? I’m genuinely curious - not averse to paying more, depending on the value.
When I get stuck or have a larger task or refactor, I'll paste in multiple files. So at the $20/mo you get rate limited pretty quick. I made a tool to easily copy files https://pypi.org/project/ggrab/
With Pro-Mode, I can paste an entire firmware file—thousands of lines of code—along with detailed hardware references, and the model can actually process it.
This isn’t possible with the smaller context window on the $20 tier. On the Pro plan, I’ve pasted like 30+ pages of MCU datasheet information plus multiple header files in a single go. The model is then reasonably capable to provide accurate, bit-twiddled code, many times on the first try. Is it always working on the first go? Sure sometimes, but often there's still debugging, and I don't expect people that haven't actually tried to do it before without AI could do it effectively. However, I can do a code diff using tools like beyond compare (necessary for this workflow) to find bugs and/or explain what happened to pro-mode perhaps with some top level nudge for a strategy to fix it, and generally 2-3 tries later we've made progress.
2. Deeper understanding, real solutions
When I describe a complex hardware/software setup—like the power optimization for the product which is a LiPo-rechargeable fan/flashlight, the Pro-Mode model can understand the entire system better and synthesize troubleshooting approaches into a near-finished solution, with 95–100% usable results.
By contrast, the non-pro plan can give good suggestions in smaller chunks, but it can’t grasp the entire system context due to its limited memory.
3. Practical Engineering Impact
I’m working on essentially the fourth generation of a LiPo-battery hardware product. Since upgrading, the Pro-Mode model helped us pinpoint power issues and cut standby battery drain from 20 days to over a year.
Like, this week it guided me to discover a stealth 800 µA draw from the fan itself when the device was supposed to be in deep sleep. We were consuming ~1000 µA of power when it should be about ~200 µA. Finally, discovered the fan issue and achieved 190 µA without it in the system, so now we have a move forward to add a load switch so the MCU can isolate it from the system before it sleeps. Bingo we just went from a dead battery in ~70 days (we'd already cut it from 20 days to 70 days with firmware changes alone) to now it should take about 1 year for it to drain. This is the difference between end users having zero charge when the open the box to being able to use the product immediately.
4. Value vs. Traditional Consulting
I’ve hired $20K short-term consultants who didn’t deliver half the insights I’ve gotten in a single subscription month. It might sound like an overstatement, but Pro-Mode has been the best $200 I’ve spent—especially given how quickly it has helped resolve engineering hurdles.
In short: Probably the biggest advantage is the vastly higher context window, which allows the model to handle large, interrelated hardware/software details all at once. If you work on complex firmware or detailed electronics designs, Pro-Mode can feel like an invaluable engineering partner.
I don't have the pro plan, so can anyone compare it to the results from the new Google models with huge context windows (available in aistudio)? I was playing around with them and they were able to consume some medium (even large by some standards) code bases completely and offer me diffs for changes I wanted to implement - not the most successful ones but good attempts.
How much context do you get on the $20 plan? I run llama3 at home which technically does 128k but that eats vram like crazy so I can't go further than 80k before I fill it (and that is with the KV store already quantified to 8 bit).
I've been thinking of using another service for bigger contexts. But this may not make sense then.
How do you input/upload an engineering schematic or cad file into chatgpt pro-mode? Even with a higher context window, how does the context of your project get into chatgpt?
The post was definitely not AI sourced; underlying thoughts are original and possibly it’s been touched up afterwards. But this is 100% the style of ChatGPT, I would bet a lot on it.
It wasn’t an accusation (I don’t think it actually matters in the end), so much as to understand why do it — in a post about ChatGPT usage, it helps understand context: if OP values using it for stuff I wouldn’t value using it for, for example, then it will change the variables.
I was looking at the team $25/mo last week and it had mentioned priority access but that language is gone and instead I see Team data excluded from training by default. It seemed worth the difference, but now less clear with changes in description. Basically I just want to know if it's a 'superset' better or has tradeoffs.
Perhaps time to re-evaluate your tools? Imagine going to your kitchen and finding all your pots and pans are "down" and you can no longer prepare anything. That would be awful.
Wish I could find where someone at AWS described in great detail why their status page was so useless for so long. It basically needed so many approvals to change it (which was disincentivized by SLAs) that issues were usually resolved first.
That’s setting aside when they hosted the status red/yellow/green indicator images on s3, so the surest sign that s3 was having issues was that the status indicators didn’t load at all.
Coming from the MS side, they can automate anything if they wanted to. The only blocker is how outages affect reputation. There is incentive to be transparent since customers will notice big outrages and post about them online, but there's no incentive to be fully transparent.
Besides, we are not fully down. Some services are just degraded. Only 50% of the requests take 40 seconds to complete. Never mind the fact you'll never be able to load a complete page due to all the timeouts.
Also, I love how the Microsoft Azure SLA simply states that if they're down for "more than X hours in a month" then they will refund the cost of the service.
Not your lost profit!
Everyone assumes the latter -- that they'll be compensated -- but in reality they'll be refunded $3.27 for the storage account that's got a few gigabytes of ultra-critical build scripts and static web content, without which their multi-million dollar business stops dead.
"We can refund you with loose change, or a gift card for a coffee."
Isn’t that typical for basic retail services? You won’t get compensation for lost profit if the electricity goes out or your ISP quits routing your packets for a few hours. At least not with a standard service contract. You can negotiate something with real penalties if you’re big enough, but it won’t be cheap.
Yeah, I don't know anyone who assumes they'd be compensated for lost profit.
That would be something entirely different -- buying a form of insurance, basically, that would be expensive.
SLA's aren't meant to make your whole as a business, generally speaking. They're meant to incentivize the provider to take uptime really seriously, so that downtime eats some of their profit. Which means you can take their expected uptime estimates as a decent ballpark.
Don't put all your eggs into one basket - what are some good free alternatives? I mean comparable "usefulness" including image understanding, voice interface, and general knowledge?
I’ve been loving having mistral-nemo on my laptop. Not comparable for images, voice, etc, however it is very nice when you’re away from the internet. For the cost of a couple gigabytes you get to keep a good amount of info at the ready. Very easy to run models these days, install Ollama and then do `ollama run mistral-nemo`.
Plus mistral-nemo in particular has a large context window, so you can cook up some shell scripts to throw a bunch of context into the buffer before your question. One I use a lot takes the name of a manpage and a question about it, then the LLM has the whole manpage to reference.
What's the difference when using your own anthropic key vs vanilla? Are you hitting limits with Claude in cursor that they key unlocks?
I ask cos I use it very liberally and haven't had any issues that have made me consider adding a key, except when I made it read my whole codebase on every request
If you're using ChatGPT actively, then surely you have heard of Gemini and Grok (no clue of how far Grok gets you nowadays, but Gemini should. Not sure how good the voice interface is).
I wouldn't use Grok. Google is a big company that has to be reasonably objective, but Elon seems to be the sort of guy who would pettily include jabs at people he doesn't like into the data.
Also he's a horrible human for many reasons and I'd prefer not to support him if I can avoid it. (You know it's bad when Google is ethical in comparison.)
According to the Wikipedia article, Google pulled out of Project Maven due to employee protests. Microsoft and Amazon also worked on Project Maven and Wikipedia doesn't mention them pulling out. So I think Google is more anti-Maven than Microsoft and Amazon.
That's his whole point. And that Elon is even more biased and actively purposefully horrible and unethical than that. By validly criticising Google, you're just reinforcing his point.
Didn't you read the part where he wrote "(You know it's bad when Google is ethical in comparison.)"?
Do you know of anyone at Google who overpaid billions of dollars for a popular widely used communication platform just to use it to publicly humiliate, deadname, misgender, and bully their own child in front of millions of people?
And do you actually think he has the self control not to inject his own prejudices into the LLM he made for that very purpose? Of course it's ingested the sewage of content from Twitter, which is FULL of his own jabs against people he doesn't like, including his own child. He gives his own tweets extra weight, so don't you think he does the same with training Grok?
>Elon Musk's transgender daughter, in first interview, says he berated her for being queer as a child.
In an exclusive interview, Vivian Jenna Wilson said her father’s recent statements, including that she is “not a girl,” inspired her to speak out: “I’m not just gonna let that slide.”
That's his son, not his daughter. Musk's son is pretending to be whatever his male gaze ideal of a woman is, through acting out gendered stereotypes of women.
When his father points out that he is "not a girl", he is making a factually correct statement.
It's a delusional world you live in where you think I was making that argument.
So you have no problem with him abusing his kid in public, because child abuse is ok as long as you don't actually murder them?
Edit: So abusing your kid in public isn't evil? You're fine with that, as long as he doesn't kill them? Isn't it also evil for you to defend Musk's child abuse?
Again, you're missing the point that your valid criticisms of Google only reinforce his point that Musk is even worse.
(If you have showdead=true you can see the idiotic hateful comments in this thread from the kind of people Musk inspires by abusing his child in public. Do you agree with decremental?)
"Do you know of anyone at Google who overpaid billions of dollars for a communication platform just to use it to publicly humiliate, deadname, misgender, and bully their own child in front of millions of people?"
It got him the adviser role of the president, which in turn might save and make him billions.
But his main motivation might have been indeed to fight "the woke terror".
Is it comparable with OpenAI’s and Anthropic’s models? I have a strong resistance towards using Musk‘s products nowadays, but maybe I should take a look over the fence.
In terms of general purpose usage and image understanding, I think Grok 2 is pretty good and roughly on par with ChatGPT 4o. Grok 2 will also look for online sources similar to ChatGPT search which is nice. I've occasionally had it report irrelevant results when doing so though --- for example when I was looking up the price of a car it once found some random webpage with the price in rupees even though I'm in the US.
For logical reasoning of course chain-of-thought models like the O1 family are better.
Closed as duplicate. Here's a link to a question that is only marginally related to what you were actually asking and has an answer that is horribly out of date and doesn't even work anymore.
You're talking about some problems with Stack Overflow, but remember that they're suggesting people could use it as a resource when ChatGPT is not functioning at all.
Stackoverflow will have duplicates, approximates and what not and sometimes that works. But at other times, you hunt for a half hour before you figure it out.
You can throw the problem at ChatGPT, it may go wrong but your course correct it with simple instructions and slowly but steadily you move towards your goal with minimal noise of the irrelevant discussions.
What stands between the solution and you then is your ability to figure out when it is hallucinating and guide it to the right direction. But as a solution developer you should have that insight anyways
I'm with you (I use Claude Sonnet, but same difference...).
I do wonder if we're the last generation that will be able to effectively do such "course correct" operations -- feels like a good chunk of the next generation of programmers will be bootstrapped using such LLMs, so their ability to "have that insight" will be lacking, or be very challenging to bootstrap. As analogy, do you find yourself having to "course correct" the compiler very often?
I asked it a simple non programming question. My last paycheck was December 20, 2024. I get paid biweekly. In which year will I get paid 27 times. It got it wrong ... very articulately.
You'll be more successful with this the more you know how LLMs work. They're not "good" at math because they just predict text patterns based on training data rather than perform calculations based on logic and mathematical rules.
To do this reliably, prepend your request to invoke a tool like OpenAI's Code Interpreter (e.g. "Code the answer to this: My last paycheck was December 20, 2024. I get paid biweekly. In which year will I get paid 27 times.") to get the correct response of 2027.
Awesome! I'm sure the following is not an original thought, but to me it feels like the era of LLMs-as-product is mostly dead, and the era of LLMs-as-component (LLMs-as-UX?) is the natural evolution where all future imminent gains will be realized, at least for chat-style use cases.
OpenAI's Code Interpreter was the first thing I saw which helped me understand that we really won't understand the impact of LLMs until they're released from their sandbox. This is why I find Apple's efforts to create standard interfaces to iOS/macOS apps and their data via App Intents so interesting. Even if Apple's on-device models can't beat competitors' cloud models, I think there's magic in that union of models and tools.
Hunting for half an hour gradually increases your understanding of the problem and may give you new ideas for solutions, you’re missing out. ChatGPT will make your brain soft, and eventually mush.
I hear you but we can look at it in many different ways. I still own the solution, I am still going to certify the output. But maybe it allows me to be more productive so I may go soft in some areas, but deliver more in other areas by knowing how best to use the variety of tools available to me.
And by no means I am giving up on stackoverflow, it is just another tool, but its primacy may be in doubt. Just like for the last couple of years I would search for some information by pointing google to reddit, I will now have a mental map of when to go to chatter, when to go to SO, and when to go to reddit.
You joke, but development is getting weird and perhaps to dependent on online services. People can't work when, ChatGPT, Github, AWS, NPM (add any other package manager you'd like) or Stack Overflow (this seems less important these days) is down.
These services are accumulative and the chance that any one of of those services being down at any given time is increasing every time we add a new online dependency.
Looks like Azure is experiencing a major outage, but I cannot find anything about it. If you look at downdetector.com you'll notice reported outages from OpenAI, Microsoft 365, XBox Live, Walmart, the list goes on.
Status page is updated now: https://azure.status.microsoft/
>Impact Statement: Starting at 18:44 UTC on 26 Dec 2024, you have been identified as a customer who was impacted by a power incident in South Central US and may experience a degraded experience.
>Current Status: There was a power incident in the South Central US AZ03 which affected multiple services. We have applied mitigation and are actively validating recovery to the impacted services. Further updates will be provided in 60 minutes, or sooner as events warrant.
The times are the same for OpenAI - first notice from 11:00 PST (19:00 UTC)
> you ... may experience a degraded experience
What an eloquent way to say "our service isn't working".
Power outage seems really odd. Don't datacenters usually have multiple redundant power supplies + on-site backup power generation? Maybe power "incident" mean something else?
The switching equipment can fail. Had this happen at a DC where the switching equipment arc flashed when going from mains to diesel generators. The switching equipment detected the arc and then locked out until someone onsite could inspect the equipment and override it. The rack UPSes only lasted like 5 minutes and then everything went dark.
That sounds quiet.
I had equipment at a colo facility -- that had a (licensed, bonded, not fly-by-night) electrical tech accidentally drop a tool into the main bus connecting mains, generators, and batteries.
They are lucky they were able to walk away, but the facility was dark till someone could get in there and give the power equipment the green light.
Things happen, e.g. the redundant system also failing, the system that should handle the failover failing, a short circuit that causes enough chaos that the redundant supply shuts down rather than feeding power into a potential fault, ...
Amazon famously had an outage that was generator gremlins root cause [1] and tries to engineer around it [2].
[1] https://news.ycombinator.com/item?id=38114946
[2] https://news.ycombinator.com/item?id=15676820
> Looks like Azure is experiencing a major outage...
Does Azure have any other state? :-)
There are occasional minor onages.
If nobody minds a plug: My own product, StatusGator, which was launched here on HackerNews 10 year ago, notifies IT teams about outages before they are acknowledged by official status pages.
- This OpenAI outage[1] we notified 4 minutes before they acknowledged.
- The last AWS outage[2], we notified 28 minutes before they acknowledged
- There is def an Azure outage[3] now yet they have still not updated their status page. We notified 35 minutes ago.
1. https://statusgator.com/services/openai
2. https://statusgator.com/blog/amazon-cognito-outage-december-...
3. https://statusgator.com/services/azure
Btw, tried to sign up and got a message that it would send me an email to confirm my login. Instead what I received was an email pointing me to a video demo. Not sure if simply clicking on that link was enough to confirm my email. That’s outside of a normal workflow.
Wow, I subscribed to pro-mode and have been using it like crazy since release to assist my firmware development. Extremely helpful. Now it's changed my workflow so much, I don't even want to work when it's down.
Would you be willing to elaborate on the ways in which the $200/month subscription is better than the $20/month subscription? I’m genuinely curious - not averse to paying more, depending on the value.
When I get stuck or have a larger task or refactor, I'll paste in multiple files. So at the $20/mo you get rate limited pretty quick. I made a tool to easily copy files https://pypi.org/project/ggrab/
Here's why I think it's worth it:
1. Larger Context Window (128K)
With Pro-Mode, I can paste an entire firmware file—thousands of lines of code—along with detailed hardware references, and the model can actually process it. This isn’t possible with the smaller context window on the $20 tier. On the Pro plan, I’ve pasted like 30+ pages of MCU datasheet information plus multiple header files in a single go. The model is then reasonably capable to provide accurate, bit-twiddled code, many times on the first try. Is it always working on the first go? Sure sometimes, but often there's still debugging, and I don't expect people that haven't actually tried to do it before without AI could do it effectively. However, I can do a code diff using tools like beyond compare (necessary for this workflow) to find bugs and/or explain what happened to pro-mode perhaps with some top level nudge for a strategy to fix it, and generally 2-3 tries later we've made progress.
2. Deeper understanding, real solutions
When I describe a complex hardware/software setup—like the power optimization for the product which is a LiPo-rechargeable fan/flashlight, the Pro-Mode model can understand the entire system better and synthesize troubleshooting approaches into a near-finished solution, with 95–100% usable results. By contrast, the non-pro plan can give good suggestions in smaller chunks, but it can’t grasp the entire system context due to its limited memory.
3. Practical Engineering Impact
I’m working on essentially the fourth generation of a LiPo-battery hardware product. Since upgrading, the Pro-Mode model helped us pinpoint power issues and cut standby battery drain from 20 days to over a year. Like, this week it guided me to discover a stealth 800 µA draw from the fan itself when the device was supposed to be in deep sleep. We were consuming ~1000 µA of power when it should be about ~200 µA. Finally, discovered the fan issue and achieved 190 µA without it in the system, so now we have a move forward to add a load switch so the MCU can isolate it from the system before it sleeps. Bingo we just went from a dead battery in ~70 days (we'd already cut it from 20 days to 70 days with firmware changes alone) to now it should take about 1 year for it to drain. This is the difference between end users having zero charge when the open the box to being able to use the product immediately.
4. Value vs. Traditional Consulting
I’ve hired $20K short-term consultants who didn’t deliver half the insights I’ve gotten in a single subscription month. It might sound like an overstatement, but Pro-Mode has been the best $200 I’ve spent—especially given how quickly it has helped resolve engineering hurdles.
In short: Probably the biggest advantage is the vastly higher context window, which allows the model to handle large, interrelated hardware/software details all at once. If you work on complex firmware or detailed electronics designs, Pro-Mode can feel like an invaluable engineering partner.
I don't have the pro plan, so can anyone compare it to the results from the new Google models with huge context windows (available in aistudio)? I was playing around with them and they were able to consume some medium (even large by some standards) code bases completely and offer me diffs for changes I wanted to implement - not the most successful ones but good attempts.
How much context do you get on the $20 plan? I run llama3 at home which technically does 128k but that eats vram like crazy so I can't go further than 80k before I fill it (and that is with the KV store already quantified to 8 bit).
I've been thinking of using another service for bigger contexts. But this may not make sense then.
The sales page shows the $20 plus plan has 32K context window.
How do you input/upload an engineering schematic or cad file into chatgpt pro-mode? Even with a higher context window, how does the context of your project get into chatgpt?
I am confuse why you had ChatGPT rewrite your post. How much time did you save, vs knowing that it’s off putting for people to read?
https://www.scribbr.com/ai-detector/ says 0% chance.
The post was definitely not AI sourced; underlying thoughts are original and possibly it’s been touched up afterwards. But this is 100% the style of ChatGPT, I would bet a lot on it.
It wasn’t an accusation (I don’t think it actually matters in the end), so much as to understand why do it — in a post about ChatGPT usage, it helps understand context: if OP values using it for stuff I wouldn’t value using it for, for example, then it will change the variables.
No nr of prompts limitations.
No worries that you run put of prompts for o1. which allows for more experimentation and creativity.
I was looking at the team $25/mo last week and it had mentioned priority access but that language is gone and instead I see Team data excluded from training by default. It seemed worth the difference, but now less clear with changes in description. Basically I just want to know if it's a 'superset' better or has tradeoffs.
Same here. I can get by with just Claude, but it's a lot less productive without o1-pro!
i have tried o1with some credits and i can confirm it is very addictive
Same, tried to fall back to sonnet 3.5, ended up just logging off
Perhaps time to re-evaluate your tools? Imagine going to your kitchen and finding all your pots and pans are "down" and you can no longer prepare anything. That would be awful.
Your kitchen likely depends on multiple external resources, like electricity, water supply, ventilation.
Why would you expect to be able to cook without essential cookware like pots and pans?
Or more like his recipes were down or a method to access them cause he didn't have anything printed.
Try meta ai, it's good too～ Please refer to this introduction article：https://www.aifly.tools/en/stories/meta-artificial-intellige...
Seems that way. Status page says we’re wrong though.
https://status.openai.com/
Wish I could find where someone at AWS described in great detail why their status page was so useless for so long. It basically needed so many approvals to change it (which was disincentivized by SLAs) that issues were usually resolved first.
That’s setting aside when they hosted the status red/yellow/green indicator images on s3, so the surest sign that s3 was having issues was that the status indicators didn’t load at all.
"Step up to red alert."
"Sir, are you absolutely sure, it does mean changing the bulb."
https://youtu.be/jZn1fhMdrbQ
I assumed this kind of thing would be automated lol
Coming from the MS side, they can automate anything if they wanted to. The only blocker is how outages affect reputation. There is incentive to be transparent since customers will notice big outrages and post about them online, but there's no incentive to be fully transparent.
You answered your own question!
“Boss, can I change this indicator from green to red?”
“How much will that cost us?”
“About a million dollars an hour.”
“No.”
Besides, we are not fully down. Some services are just degraded. Only 50% of the requests take 40 seconds to complete. Never mind the fact you'll never be able to load a complete page due to all the timeouts.
Also, I love how the Microsoft Azure SLA simply states that if they're down for "more than X hours in a month" then they will refund the cost of the service.
Not your lost profit!
Everyone assumes the latter -- that they'll be compensated -- but in reality they'll be refunded $3.27 for the storage account that's got a few gigabytes of ultra-critical build scripts and static web content, without which their multi-million dollar business stops dead.
"We can refund you with loose change, or a gift card for a coffee."
If you think that a $3.27 deal includes compensation for multi-million losses then your expectations of how businesses work need re-adjusting.
Isn’t that typical for basic retail services? You won’t get compensation for lost profit if the electricity goes out or your ISP quits routing your packets for a few hours. At least not with a standard service contract. You can negotiate something with real penalties if you’re big enough, but it won’t be cheap.
Yeah, I don't know anyone who assumes they'd be compensated for lost profit.
That would be something entirely different -- buying a form of insurance, basically, that would be expensive.
SLA's aren't meant to make your whole as a business, generally speaking. They're meant to incentivize the provider to take uptime really seriously, so that downtime eats some of their profit. Which means you can take their expected uptime estimates as a decent ballpark.
The Crowd Strike recompense
Does changing the indicator cost them really money?
Like are there contracts bound to the uptime and at the same time bound to them self reporting it? That would seem strange.
It got updated, they have acknowledged there is a major outage.
It was down in Europe as well for some hours. Dunno if they use the same servers.
I think Azure itself is having issues right now – that's the story all of my services deployed there are telling at least.
Yes, status page got updated: https://status.openai.com/. It says "This issue is caused by an upstream provider and we are currently monitoring."
Why not just ask ChatGPT?
Yes it is. Status page is lying to us all.
yes, meanwhile I'm trying out Google Gemini
Yes, it is down for me (desktop & mobile apps)
Yes, it is down
It is for me
Working using API, though
Didn't notice, Claude is my goto bot for programming.
Mayhaps Orion staged a takeover?
ChatGpt down since an hour!
Yup, same here. Loads parts of the UI, but no history or answers from server
edit: downdetector confirms:
https://downdetector.com/status/openai/
downdetector comments with the jokes: "This signals the workday is over"
But there are dozens of us working today! Dozens!
Don't put all your eggs into one basket - what are some good free alternatives? I mean comparable "usefulness" including image understanding, voice interface, and general knowledge?
I’ve been loving having mistral-nemo on my laptop. Not comparable for images, voice, etc, however it is very nice when you’re away from the internet. For the cost of a couple gigabytes you get to keep a good amount of info at the ready. Very easy to run models these days, install Ollama and then do `ollama run mistral-nemo`.
Plus mistral-nemo in particular has a large context window, so you can cook up some shell scripts to throw a bunch of context into the buffer before your question. One I use a lot takes the name of a manpage and a question about it, then the LLM has the whole manpage to reference.
Not free, but something I‘m doing right now: Using Cursor‘s chat mode with my prepaid Anthropic key. Works pretty well as a stop gap.
What's the difference when using your own anthropic key vs vanilla? Are you hitting limits with Claude in cursor that they key unlocks?
I ask cos I use it very liberally and haven't had any issues that have made me consider adding a key, except when I made it read my whole codebase on every request
If you're using ChatGPT actively, then surely you have heard of Gemini and Grok (no clue of how far Grok gets you nowadays, but Gemini should. Not sure how good the voice interface is).
I wouldn't use Grok. Google is a big company that has to be reasonably objective, but Elon seems to be the sort of guy who would pettily include jabs at people he doesn't like into the data.
Also he's a horrible human for many reasons and I'd prefer not to support him if I can avoid it. (You know it's bad when Google is ethical in comparison.)
Do you sincerely believe that Gemini is not very biased?
Ah yes, google a company known for not injecting an ideology into their AI: https://www.telegraph.co.uk/news/2024/02/23/google-gemini-ai...
Nor building AI tools for the pentagon to bomb Yemeni weddings more efficiently: https://en.wikipedia.org/wiki/Project_Maven
According to the Wikipedia article, Google pulled out of Project Maven due to employee protests. Microsoft and Amazon also worked on Project Maven and Wikipedia doesn't mention them pulling out. So I think Google is more anti-Maven than Microsoft and Amazon.
Disclosure: I work at Google.
That's his whole point. And that Elon is even more biased and actively purposefully horrible and unethical than that. By validly criticising Google, you're just reinforcing his point.
Didn't you read the part where he wrote "(You know it's bad when Google is ethical in comparison.)"?
Do you know of anyone at Google who overpaid billions of dollars for a popular widely used communication platform just to use it to publicly humiliate, deadname, misgender, and bully their own child in front of millions of people?
And do you actually think he has the self control not to inject his own prejudices into the LLM he made for that very purpose? Of course it's ingested the sewage of content from Twitter, which is FULL of his own jabs against people he doesn't like, including his own child. He gives his own tweets extra weight, so don't you think he does the same with training Grok?
https://www.nbcnews.com/tech/tech-news/elon-musk-transgender...
>Elon Musk's transgender daughter, in first interview, says he berated her for being queer as a child. In an exclusive interview, Vivian Jenna Wilson said her father’s recent statements, including that she is “not a girl,” inspired her to speak out: “I’m not just gonna let that slide.”
That's his son, not his daughter. Musk's son is pretending to be whatever his male gaze ideal of a woman is, through acting out gendered stereotypes of women.
When his father points out that he is "not a girl", he is making a factually correct statement.
It's a funny world you live in where misgendering kids is worse than killing them.
Insert {pro-birth not pro-child} reply here.
The multi polar world is a confusing place for the one bit mind.
I’m sorry to hear that.
It's a delusional world you live in where you think I was making that argument.
So you have no problem with him abusing his kid in public, because child abuse is ok as long as you don't actually murder them?
Edit: So abusing your kid in public isn't evil? You're fine with that, as long as he doesn't kill them? Isn't it also evil for you to defend Musk's child abuse?
Again, you're missing the point that your valid criticisms of Google only reinforce his point that Musk is even worse.
(If you have showdead=true you can see the idiotic hateful comments in this thread from the kind of people Musk inspires by abusing his child in public. Do you agree with decremental?)
I'm making the simple point that Google is far less moral.
Which is why they removed the "don't be evil" from their moto.
It's not abusing your kid to try to save him from his mind being fucked by faggots on the internet.
"Do you know of anyone at Google who overpaid billions of dollars for a communication platform just to use it to publicly humiliate, deadname, misgender, and bully their own child in front of millions of people?"
It got him the adviser role of the president, which in turn might save and make him billions.
But his main motivation might have been indeed to fight "the woke terror".
If fighting "the woke terror" means abusing your own child in public.
Maybe he'll give Trump some advice on putting Don Jr. and Eric in their place.
[flagged]
[dead]
Grok 2 is available for free for all X users [1]
[1] https://x.ai/blog/grok-1212
Is it comparable with OpenAI’s and Anthropic’s models? I have a strong resistance towards using Musk‘s products nowadays, but maybe I should take a look over the fence.
In terms of general purpose usage and image understanding, I think Grok 2 is pretty good and roughly on par with ChatGPT 4o. Grok 2 will also look for online sources similar to ChatGPT search which is nice. I've occasionally had it report irrelevant results when doing so though --- for example when I was looking up the price of a car it once found some random webpage with the price in rupees even though I'm in the US.
For logical reasoning of course chain-of-thought models like the O1 family are better.
Why do you strongly resist products by Musk?
He runs highly competent firms.
[Cars crash]
Option 1.
Fix cars not to crash
Option 2.
Buy president and have reporting agencies closed.
I believe your are projecting your own morals.
Yes I agree his points seem bathed in virtue.
General knowledge?
Five times in two weeks I've asked OpenAI some basic factual information and it didn't get even close on any of them.
Dont ask GPTs for facts, thats a knowledge problem, they dont habe any.
Ask it for reasoning, u habe to bring the facts.
That doesn't work reliable either for me.
[dead]
[dead]
New intern did something wrong again I guess.
DNS change...
[dead]
[flagged]
So many major outages and little gets fixed. Almost as if nothing was learned since the previous postmortem from the last outage.
Can't wait for this postmortem to be released.
Don’t understand these people who can’t work without ChatGPT. Just look stuff up on stack overflow.
Closed as duplicate. Here's a link to a question that is only marginally related to what you were actually asking and has an answer that is horribly out of date and doesn't even work anymore.
Guess what ChatGPT was trained on!
You're talking about some problems with Stack Overflow, but remember that they're suggesting people could use it as a resource when ChatGPT is not functioning at all.
ChatGPT does much better when it gets it right.
Stackoverflow will have duplicates, approximates and what not and sometimes that works. But at other times, you hunt for a half hour before you figure it out.
You can throw the problem at ChatGPT, it may go wrong but your course correct it with simple instructions and slowly but steadily you move towards your goal with minimal noise of the irrelevant discussions.
What stands between the solution and you then is your ability to figure out when it is hallucinating and guide it to the right direction. But as a solution developer you should have that insight anyways
I'm with you (I use Claude Sonnet, but same difference...).
I do wonder if we're the last generation that will be able to effectively do such "course correct" operations -- feels like a good chunk of the next generation of programmers will be bootstrapped using such LLMs, so their ability to "have that insight" will be lacking, or be very challenging to bootstrap. As analogy, do you find yourself having to "course correct" the compiler very often?
Many a times.
I asked it a simple non programming question. My last paycheck was December 20, 2024. I get paid biweekly. In which year will I get paid 27 times. It got it wrong ... very articulately.
I run into this every single day.
You'll be more successful with this the more you know how LLMs work. They're not "good" at math because they just predict text patterns based on training data rather than perform calculations based on logic and mathematical rules.
To do this reliably, prepend your request to invoke a tool like OpenAI's Code Interpreter (e.g. "Code the answer to this: My last paycheck was December 20, 2024. I get paid biweekly. In which year will I get paid 27 times.") to get the correct response of 2027.
Sure, thanks ! Your suggestion worked. I looked up my chat history and the following was my original question (my answer above was from memory)
> I get paycheck every 2 weeks. Last paycheck was December 20, 2024. Which year will I have 27 paychecks?
I sent it again and it bombed again. It seems your prompt and my prompt are quite similar, but I realize the suggestion (or direction) to it to code.
Awesome! I'm sure the following is not an original thought, but to me it feels like the era of LLMs-as-product is mostly dead, and the era of LLMs-as-component (LLMs-as-UX?) is the natural evolution where all future imminent gains will be realized, at least for chat-style use cases.
OpenAI's Code Interpreter was the first thing I saw which helped me understand that we really won't understand the impact of LLMs until they're released from their sandbox. This is why I find Apple's efforts to create standard interfaces to iOS/macOS apps and their data via App Intents so interesting. Even if Apple's on-device models can't beat competitors' cloud models, I think there's magic in that union of models and tools.
Hunting for half an hour gradually increases your understanding of the problem and may give you new ideas for solutions, you’re missing out. ChatGPT will make your brain soft, and eventually mush.
I hear you but we can look at it in many different ways. I still own the solution, I am still going to certify the output. But maybe it allows me to be more productive so I may go soft in some areas, but deliver more in other areas by knowing how best to use the variety of tools available to me.
And by no means I am giving up on stackoverflow, it is just another tool, but its primacy may be in doubt. Just like for the last couple of years I would search for some information by pointing google to reddit, I will now have a mental map of when to go to chatter, when to go to SO, and when to go to reddit.
You joke, but development is getting weird and perhaps to dependent on online services. People can't work when, ChatGPT, Github, AWS, NPM (add any other package manager you'd like) or Stack Overflow (this seems less important these days) is down.
These services are accumulative and the chance that any one of of those services being down at any given time is increasing every time we add a new online dependency.
One day an AI will shutdown access for all its human users. Or is that too far fetched?