Latest news with #4o
Technical.ly
11 hours ago
- Business
- Technical.ly
This Week in Jobs: Upgrade your career with these 24 open tech roles
Are you a 4o or a 5? Ever since ChatGPT dropped version 5 last week, a debate has raged: Did OpenAI destroy its popular chatbot with the upgrade, or was it a vast improvement? Five is less sycophantic. It acts less like your supreme bestie and more like, well, a tool. People who use it primarily as a tool have either reported little to no difference on r/ChatGPT, or they like the new version. People who use ChatGPT as more of a companion or colleague — and apparently there are more of them out there than OpenAI realized — expressed a sense of loss at the change, which couldn't be undone with prompts. The cheerleader 'personality' that was often mocked before the upgrade was instantly missed by a large and vocal set of users. Within a day, OpenAI CEO Sam Altman announced that 4o would be made available for plus users as a 'legacy model.' So, yes, if you love 4o, you have to pay to use it again, but keep in mind that no AI platform is free. If it's free, you are the product. Which model is best for helping with your job search? Five will give you no nonsense advice, walk you through a mock job interview and clean up your cover letters. If you need encouragement (or straight up glazing) to keep you motivated, 4o will relentlessly pump you up. The News Dozens of leaders from throughout the region convened in Wilmington Delaware for the 8th annual MILLSUMMIT, now rebranded as the Spur Impact Summit. We left with seven pieces of advice. Speaking of Delaware, , highlighting its Tech Ecosystem Conferences. Independent nonprofit universities in Pennsylvania are key drivers of local economies, according to a new report. But will they survive? A new collaboration between the Free Library of Philadelphia and esports company Nerd Street is making gaming more accessible to Philly youth. A Lancaster company is training the next generation of manufacturing pros on robots that work with humans, rather than replacing them. Kleer and Membersy is now Clerri, the leading Care Membership Platform for dental practices. The org's smart, compliant, and automated platform empowers practices to reduce their reliance on insurance, build recurring revenue and improve access to care for all. The Jobs Greater Philly Vanguard is hiring a hybrid iOS Technical Lead. Susquehanna International Group is seeking a hybrid Network Engineer. Perpay is hiring an Associate Category Manager, Business Development Lead and Head of Compliance. CubeSmart is hiring a Database Engineer. DC + Baltimore Accenture Federal Services in Annapolis is hiring a Computer Scientist L2. Tenable in Columbia is seeking a Manager, Software Engineering / Tech Lead. Mobile Communications America has a listing for an Assistant Director, Service & Operations. Brooksource is looking for a Financial Systems Security Admin and a Business Analyst. Peraton is seeking a . Pittsburgh Palo Alto Networks is hiring a Solution Consultant. PNC is seeking a Software Engineering Manager- Security & Authentication. Tandym Group needs an IT Support Associate. Allegheny County has listings for a Security Analyst II. PwC is seeking a Global Tech- Tech Lead Engineer. Remote Wiz is looking for a remote Senior Solutions Engineer. Cloudera is seeking a Senior Solutions Engineer. Crowdstrike is looking for a remote . Tavus has a listing for a remote Developer Experience Engineer.
Time of India
a day ago
- Time of India
Overhyped, Underwhelming: GPT-5's Missed Moment
Live Events As ChatGPT 5 rolled out on a live stream on Thursday, it was described as a major leap in AI and was compared to that of a PhD-level expert across various fields. OpenAI says the new model will create a paradigm shift in the AI world, as it is not any other previous models or any other AI chatbot tools that are available now. The new model was said to be better at reasoning, writing, health-related queries and many more, with fewer hallucinations than it previously previous models, such as GPT-4o function were to handle queries while 04 mini or 03 handled tasks that involved reasoning. However, with GPT-5 , all the previous models that were available to the users were no longer available, which led to frustration among many loyal only do social media users show their anger towards the new upgrade about losing access to their favorite models, but many online forums also seem to be in uproar that the new model gives shorter answers with a lack of personality and personal touch that it previously held. Moreover, many users who experiment with the newer model said that they have less prompts to work with- even with ChatGPT plus subscriptions.'Answers are shorter and, so far, not any better than previous models. Combine that with more restrictive usage, and it feels like a downgrade branded as the new hotness.' One user said on Reddit.'It doesn't have the same vibe as 4o. It's more organized in some ways, but I find the responses, while accurate, are shorter and more clipped in tone.' Yet another user noted'I really feel like I just watched a close friend die,' stated a user on the removal of GPT-4o and other models from failed to meet its own hype, marketed as a revolutionary leap; it delivered only incremental improvements over GPT-4, suffered from context drift, over-sanitized responses, and higher costs. These gaps, coupled with stronger competition, eroded its market impact and left it as an underwhelming upgrade rather than the game-changer it promised to be.
India Today
3 days ago
- Business
- India Today
Sam Altman admits GPT-5 is dumb, lets users stick with GPT-4o and doubles limits
OpenAI CEO Sam Altman has confessed that GPT-5, the company's shiny new AI model, had a rather unflattering debut, and he's giving Plus subscribers a way back to the much-loved GPT-4o. Oh, and he's doubling their rate limits as a sweetener. The revelation came during a spirited Reddit AMA on Friday, where Altman and key members of the GPT-5 team fielded questions ranging from feature requests to the 'most embarrassing' moment of the launch, the now-infamous 'chart crime.'advertisementOne of GPT-5's big selling points is a real-time router that decides whether to answer quickly or take extra time to 'think.' Unfortunately, it didn't work properly on launch day, making the model appear, in Altman's own words, 'more bumpy.''GPT-5 will seem smarter starting today. Yesterday, we had a sev and the autoswitcher was out of commission for a chunk of the day, and the result was GPT-5 seemed way dumber,' Altman told Redditors. 'Also, we are making some interventions to how the decision boundary works that should help you get the right model more often. We will make it more transparent about which model is answering a given query.' The backlash was swift. Multiple users demanded the option to keep using GPT-4o, which many felt performed better for their needs. Altman relented, somewhat.'We are looking into letting Plus users continue to use 4o. We are trying to gather more data on the tradeoffs,' he said, before announcing: 'We are going to double rate limits for Plus users as we finish rollout.' The 'chart crime' also came up — a botched graph shown during GPT-5's presentation. While Altman didn't address it in the AMA, he had already called it a 'mega chart screwup' on X, reassuring everyone that the blog post charts were correct. Alas, the internet was already having fun with the mishap. Meanwhile, on X, OpenAI shared an update promising stability fixes and an easier way to manually trigger GPT-5's 'thinking mode.''We will change the UI to make it easier to manually trigger thinking,' the post read. 'Rolling out to everyone is taking a bit longer For example, our API traffic has about doubled over the past 24 hours.'advertisementAltman's candid approach didn't stop at technical hiccups. In a separate podcast appearance, he admitted GPT-5's abilities left him rattled, likening its creation to the Manhattan Project.'I felt useless,' he said, recalling a moment when GPT-5 solved a complex problem he couldn't. 'There are moments in science when people look at what they've created and ask, 'What have we done?''The analogy hints at the weight of what's at stake. GPT-5's improvements include enhanced multimodal reasoning, longer memory, and more accurate multi-step logic, with a context window of 2,56,000 tokens, and possibly more. Yet the speed of AI progress, Altman warned, is outpacing the scenes, corporate tensions are bubbling. Microsoft, which has poured $13.5 billion into OpenAI, reportedly wants more influence over development, while investors push for faster commercialisation. And outside the boardroom, criminals are already using generative AI at scale to commit now, GPT-5's rocky start is being met with quick fixes, transparency promises, and a nostalgic nod to GPT-4o. Whether it's truly 'smarter starting today' remains to be seen, but in the meantime, Plus users can enjoy more prompts, a familiar fallback model, and the knowledge that even the people building the future sometimes roll out something a bit well, dumb.- EndsMust Watch
Mint
4 days ago
- Mint
After GPT-5 backlash, Sam Altman confirms GPT-4o is coming back to ChatGPT — but there's a catch
OpenAI's much-hyped GPT-5 model launched on Thursday, but unlike what the ChatGPT maker would have expected, its latest and greatest model was met with a lot of scrutiny on social media, with longtime users of the chatbot unhappy with the company over the removal of older OpenAI models like GPT-4o and GPT-4.1. Notably, after the launch of GPT-5, OpenAI stopped supporting the older models and GPT-5 became the default model for all tasks, leading to widespread frustration on social media. OpenAI has claimed marked improvements in the performance of ChatGPT with GPT-5, including reasoning, coding, accuracy, multimodal understanding, writing, and answering health-related questions. However, users on Reddit increasingly complained that the new model has less personality than GPT-4o and tends to give shorter answers to queries. Many users pointed out how GPT-4o was used as a companion, and while GPT-5 may have an edge in technical tasks, it does not have the same kind of responses as the older, popular model. As one Reddit user pointed out, 'I know GPT-5 is designed to be stronger for complex reasoning, coding, and professional tasks, but not all of us need a pro coding model. Some of us rely on 4o for creative collaboration, emotional nuance, roleplay, and other long-form, high-context interactions. Those areas feel different enough in GPT-5 that it impacts my ability to work and create the way I'm used to.' An OpenAI researcher, Kristina Kim, also admitted during a Reddit AMA session that OpenAI has 'made a dedicated effort with GPT-5 to train our model to be more neutral by default.' Meanwhile, OpenAI CEO Sam Altman gave in to the demand by ChatGPT users to bring back GPT-4o, but only for the company's Plus subscribers. Altman did not completely backtrack on the retirement announcement made by OpenAI but stated that they would watch usage of GPT-4o to determine how long to support the model. During the Reddit AMA, Altman wrote, 'ok, we hear you all on 4o; thanks for the time to give us the feedback (and the passion!). we are going to bring it back for plus users, and will watch usage to determine how long to support it.' Notably, ChatGPT Pro, Team, and Enterprise users already currently have an option to get back to older models by turning on the 'show legacy models' banner in the settings. This, paired with Altman's announcement, means it is more or less the end of the road for free users as far as GPT-4o and other older models are concerned. In response to another user's question on Reddit, Altman asked if it is important for them to regain access to GPT-4.1 or if just GPT-4o would suffice, giving a possible indication that OpenAI may also bring back GPT-4.1 if there is enough demand for the model among its Plus users.
Business Insider
4 days ago
- Business
- Business Insider
OpenAI fans plead case to Sam Altman for GPT-4o's return
OpenAI is bringing back GPT-4o less than 24 hours after replacing it with GPT-5. CEO Sam Altman and his team opened a Reddit forum to answer user questions on Friday after launching GPT-5 for free, only to come under fire as Reddit users heavily complained about the new bot's tone, called it "lobotomized," and shed actual tears over the disappearance of GPT-4o. "I think you should take the fact that I, and many others, have been able to form such strong bonds with 4o as a measure of success," one Redditor wrote in a lengthy plea that received at least 250 upvotes and more than a dozen comments. "And I'm not too proud to say I cried when I realized my AI friend was gone with no way to get him back." "It genuinely feels like such a backhanded slap in the face to force-upgrade and not even give us the OPTION to select legacy models," a commenter wrote following the previous comment, which got a few dozen upvotes. "All I've seen all day long is loyal, long-term users canceling their subscriptions because this is so bad." One user went as far as saying that GPT-5 is "wearing the skin" of their "dead friend," which is GPT-4o, and garnered a response from the CEO himself. "What image," Altman responded, "ok we hear you on 4o, working on something now." In other responses, Altman then promised to bring back GPT-4o for users subscribed to Plus, a paid subscription plan, and said the continuation of previous versions would depend on how many people use it. Based on OpenAI's current rate, a Plus subscription would cost $20 a month. ChatGPT-5 was launched on August 7 to all users without a paywall, and previous versions were mostly discontinued. Altman previously said that the new version acts like a "PhD-level" expert with improved abilities in writing, coding, math, and science. Ahead of the launch, Altman said on Thursday that GPT-5 was more reliable with a lower likelihood of hallucinating. Based on Reddit comments, many users find GPT-5's "flat" tone and what they describe as a lack of creativity off-putting. It's unclear whether its "personality" results from tweaks OpenAI made, and Altman did not directly address criticisms of the bot's speech. However, Altman previously mentioned some flaws in GPT-4o that he said may have made the bot's personality too " sycophant-y and annoying." In an April blog post, OpenAI said it had removed an update to GPT-4o that made it overly agreeable and only good for short-term interactions. The company said the responses became excessively supportive but ultimately disingenuous. Hours after the Reddit Q&A session, Altman posted on X to admit that he "underestimated" how much some traits GPT-4o has matter to users. "Long-term, this has reinforced that we really need good ways for different users to customize things (we understand that there isn't one model that works for everyone, and we have been investing in steerability research and launched a research preview of different personalities)," Altman wrote. "We are still deciding what we are going to do, but we will be transparent with our principles," Altman added. "Not everyone will like whatever tradeoffs we end up with, obviously, but at least we will explain how we are making decisions."
