OpenAI, the creator of ChatGPT, is facing significant backlash from industry experts who are disappointed with the latest GPT-4 model.
OpenAI, the creator of ChatGPT, is facing significant backlash from industry experts who
are disappointed with the latest GPT-4 model.
Once praised for its speed, the recent version of the AI model has been widely criticised
for its perceived weakness, laziness, and lack of intelligence. Despite OpenAI’s initial
claims, users have expressed their frustration with the model, deeming it subpar
compared to its predecessors.
Many users have taken to Twitter and online forums to voice their concerns about GPT-
4’s underwhelming performance. They report instances of incorrect responses, an
apparent loss of track of input data, and difficulties in following instructions. Even the
simple task of including brackets in software code has proven to be a challenge for the
AI model. Essentially, GPT-4 appears to be merely regurgitating recently generated prompts.
According to one developer, GPT-4’s current state is highly disappointing, as it cannot be
relied upon for adding code functions to websites. Comparing it to a luxury car that has
been transformed into an old vehicle, the developer stated that they would never pay for
a service like this.
Faster But Less Useful?
Another user highlighted the issue of deteriorating quality despite faster output
generation. Basic queries, such as improving writing clarity or generating ideas, have
become difficult tasks. The declining quality raises concerns about the model’s
usefulness.
Observers have noted the AI model’s tendency to loop outputs haphazardly, leading to
the conclusion that it has become lost its previous usefulness. This revelation is
particularly shocking considering the significant advancements and popularity that AI
has achieved in a short period.
Experts have also suggested that a radical redesign of GPT-4 might be the reason
behind its diminished performance. Previously known for its accuracy and high cost, the
model’s transition to faster but less accurate results has left many disappointed.
The launch of the larger GPT-4 in March was highly anticipated in the tech sector. Its
multi-modal functionality, enabling comprehension of both text and images, further
added to the excitement. However, it appears that such models may not be long-lasting
favourites among developers.
Reduced Costs Might Lead to Reduced Quality
Industry insiders speculate that OpenAI may be pursuing a Mixture of Experts (MOE)
approach, wherein smaller GPT-4 models specialise in various subject areas. This
approach aims to reduce costs while maintaining quality. However, the transition to this
new system seems to have affected GPT-4’s performance, leaving users dissatisfied.
OpenAI has yet to provide an official response to these concerns. As leaked details of
GPT-4’s architecture circulate online, it remains to be seen how the company will
address the issues raised by users and experts alike.
In a Nutshell
OpenAI’s highly anticipated GPT-4 model has faced criticism for its degraded
performance and weakened logic. Users have expressed disappointment with its
accuracy, reliability, and ability to follow instructions. While a radical redesign and the
adoption of a Mixture of Experts approach may have caused these issues, OpenAI’s
response to the situation remains unclear. The future of GPT-4 and its place among
developers’ preferences is uncertain at this point.
