Some javascripts cause an error in ChatGPT web after AUG 3 update

-edit:

I’ve just run into another error with a question about a script. I’ve never had these issues before and now I’m getting errors like this depending on the script I paste into the prompt. Anyone else had this problem?

The following prompt causes an error no matter how often I try it. I send prompts like this all the time and never encounter any issues like this, but there must be something about this particular prompt that bugs it out:

this change doesn't fix it either. Let's examine the steps together:
1: When I select a category and brand, We got this xhr payload:
action: fetch_model_repairs
selectedCategory: Smartphone
selectedBrand: Apple

Which returns a response like this:
[
    {
        "model_fk_id": "1",
        "repair_fk_id": "1",
        "ro_fk_id": "2",
        "price": "0.00",
        "price_source": "1",
        "duration": "0",
        "duration_source": "1",
        "duration_unit": "1",
        "repair_active": "1",
        "m_name": "iPhone 14",
        "c_name": "Smartphone",
        "b_name": "Apple",
        "repair_name": "Screen",
        "option_name": "Original"
    }
]

And renders the table with this data.

When we perform a model search, we send this xhr payload:
action: fetch_model_repairs
selectedCategory: Select a category
selectedBrand: Select a brand
modelIds[]: 1

which returns this response:
[
    {
        "model_fk_id": "1",
        "repair_fk_id": "1",
        "ro_fk_id": "2",
        "price": "0.00",
        "price_source": "1",
        "duration": "0",
        "duration_source": "1",
        "duration_unit": "1",
        "repair_active": "1",
        "m_name": "iPhone 14",
        "c_name": "Smartphone",
        "b_name": "Apple",
        "repair_name": "Screen",
        "option_name": "Original"
    }
]

But the table isn't rendered. Why is the table rendered correctly for the first action, but not for the second? I think the issue is in here somewhere:
    // Fetch Model Repairs function
    function fetchModelRepairs(modelIds) {
        var selectedCategory = $('#category-selection').val();
        var selectedBrand = $('#brand-selection').val();
    
        // Do not send the AJAX request if either the category or brand is 'All' and no model IDs are provided
        if ((selectedCategory === 'All' || selectedBrand === 'All') && !modelIds) {
            return;
        }
    
        var data = {
            'action': 'fetch_model_repairs',
            'selectedCategory': selectedCategory,
            'selectedBrand': selectedBrand
        };
    
        var selectedRepairType = $('#repair-type-selection').val();
        // If a specific repair type is selected, add it to the data sent to the server
        if (selectedRepairType !== 'All') {
            data.selectedRepairType = selectedRepairType;
        }
    
        // If model IDs are provided, add them to the data sent to the server
        if (modelIds) {
            data.modelIds = modelIds;
        }
    
        $.ajax({
            url: ajaxurl,
            type: 'POST',
            data: data,
            dataType: 'json',
            success: function(data) {
                // Create a new array to hold the structured data
                var structuredData = [];
                // Iterate over the fetched data
                data.forEach(function(row) {
                    // Check if an object for the current model and repair type already exists
                    var existingEntry = structuredData.find(function(entry) {
                        return entry.c_name === row.c_name && entry.b_name === row.b_name && entry.m_name === row.m_name && entry.repair_name === row.repair_name;
                    });
                    // If it exists, add a new repair option to the options array
                    if (existingEntry) {
                        existingEntry.repair_options.push({
                            option_name: row.option_name,
                            price: row.price,
                            duration: row.duration,
                            duration_unit: row.duration_unit,
                            repair_active: row.repair_active,
                            model_fk_id: row.model_fk_id,
                            repair_fk_id: row.repair_fk_id,
                            ro_fk_id: row.ro_fk_id
                        });
                    } else {
                        // If it doesn't exist, create a new object and add it to the array
                        structuredData.push({
                            c_name: row.c_name,
                            b_name: row.b_name,
                            m_name: row.m_name,
                            repair_name: row.repair_name,
                            repair_options: [{
                                option_name: row.option_name,
                                price: row.price,
                                duration: row.duration,
                                duration_unit: row.duration_unit,
                                repair_active: row.repair_active,
                                model_fk_id: row.model_fk_id,
                                repair_fk_id: row.repair_fk_id,
                                ro_fk_id: row.ro_fk_id
                            }]
                        });
                    }
                });
                // Replace the fetched data with the structured data
                fetchedData = structuredData;
    
                renderTable(); // Render the table using fetched data
            },
            error: function(jqXHR, textStatus, errorThrown) {
                console.log('Error: ' + textStatus + ' ' + errorThrown);
            }
        }); 
    }

    
    // New repairModelSearch function
    function repairModelSearch(searchTerm) {
        $.ajax({
            url: ajaxurl,
            type: 'POST',
            data: {
                'action': 'repair_model_search',
                'searchTerm': searchTerm
            },
            dataType: 'json',
            success: function(data) {
                // data should be an array of model IDs
                // Call fetchModelRepairs with the fetched model IDs
                fetchModelRepairs(data);
                renderTable(); // Render the table using fetched data
            },
            error: function(jqXHR, textStatus, errorThrown) {
                console.log('Error: ' + textStatus + ' ' + errorThrown);
            }
        }); 
    }    




    function renderTable() {
        var selectedCategory = $('#category-selection').val();
        var selectedBrand = $('#brand-selection').val();
        var selectedRepairType = $('#repair-type-selection').val();
      
        var data = fetchedData.filter(function(row) {
            var categoryMatches = selectedCategory === 'All' || row.c_name === selectedCategory;
            var brandMatches = selectedBrand === 'All' || row.b_name === selectedBrand;
            var repairTypeMatches = selectedRepairType === 'All' || row.repair_name === selectedRepairType;
            return categoryMatches && brandMatches && repairTypeMatches;
        });
    
        var repairOptions = getUniqueRepairOptions(data);
    
        var thead = $('#model-repair-pricing table thead tr');
        var tbody = $('#model-repair-pricing table tbody');
    
        // Clear the tbody and remove old repair options from thead
        tbody.html('');
        thead.find('.repair-option-th').remove();
    
        // Add new repair options to thead
        repairOptions.forEach(function(option) {
            thead.append(`<th class="repair-option-th">${option}</th>`);
        });
    
        data.forEach(function(row) {
            var html = `
                <tr>
                    <td class="cat-row">${row.c_name}</td>
                    <td class="brand-row">${row.b_name}</td>
                    <td class="model-repairs">${row.m_name}</td>
                    <td class="repair-row">${row.repair_name}</td>`;
    
            repairOptions.forEach(function(option) {
                var repairOption = row.repair_options.find(function(ro) { return ro.option_name === option; });
                if (repairOption) {
                    html += `
                        <td class="repair-option-td">
                            <div class="field-row">
                                <span class="field-label">Price:</span> <input type="text" class="price-input field-input" data-model-repair-id="${repairOption.model_fk_id}-${repairOption.repair_fk_id}-${repairOption.ro_fk_id}" value="${repairOption.price}">
                                <span class="field-label">Active:</span> <input type="checkbox" class="active-checkbox" data-model-repair-id="${repairOption.model_fk_id}-${repairOption.repair_fk_id}-${repairOption.ro_fk_id}"${repairOption.repair_active === '1' ? ' checked' : ''}>
                            </div>
                            <div class="field-row">
                                <span class="field-label">Time:</span> <input type="text" class="time-input field-input" data-model-repair-id="${repairOption.model_fk_id}-${repairOption.repair_fk_id}-${repairOption.ro_fk_id}" value="${repairOption.duration}">
                                <select class="unit-select" data-model-repair-id="${repairOption.model_fk_id}-${repairOption.repair_fk_id}-${repairOption.ro_fk_id}">
                                    <option value="1"${repairOption.duration_unit == '1' ? ' selected' : ''}>Minutes</option>
                                    <option value="2"${repairOption.duration_unit == '2' ? ' selected' : ''}>Hours</option>
                                    <option value="3"${repairOption.duration_unit == '3' ? ' selected' : ''}>Days</option>
                                </select>
                            </div>
                        </td>`;
                } else {
                    html += '<td class="repair-option-td"></td>';
                }
            });
    
            html += `</tr>`;
            tbody.append(html);
        });
    
        // Only display the switches div if there are repair options
        if (repairOptions.length > 0) {
            $('#switches').css('display', 'block');
            $('#switches-toggle').css('display', 'block');
            
        }
    
        initializeColumnSwitches(); // Initialize the column switches here
    
        bindDropdownChangeEvents(); 
    }

We're getting the data but we don't trigger the rendertable function correctly or something like that. What do you think causes this?

The console log prints:
POST https://chat.openai.com/backend-api/conversation 403
(anonymous) @ _app-695c0ff86878a9c6.js:28
(anonymous) @ _app-695c0ff86878a9c6.js:33
Z @ _app-695c0ff86878a9c6.js:33
tb @ _app-695c0ff86878a9c6.js:33
(anonymous) @ _app-695c0ff86878a9c6.js:33
td @ _app-695c0ff86878a9c6.js:33
(anonymous) @ _app-695c0ff86878a9c6.js:33
(anonymous) @ main-664b20fa39df3103.js:1
(anonymous) @ main-664b20fa39df3103.js:1
n @ main-664b20fa39df3103.js:1
u @ main-664b20fa39df3103.js:1
(anonymous) @ main-664b20fa39df3103.js:1
(anonymous) @ main-664b20fa39df3103.js:1
(anonymous) @ _app-695c0ff86878a9c6.js:33
(anonymous) @ _app-695c0ff86878a9c6.js:33
(anonymous) @ main-664b20fa39df3103.js:1
(anonymous) @ main-664b20fa39df3103.js:1
n @ main-664b20fa39df3103.js:1
u @ main-664b20fa39df3103.js:1
(anonymous) @ main-664b20fa39df3103.js:1
(anonymous) @ main-664b20fa39df3103.js:1
U.publicApiCompletionStream @ _app-695c0ff86878a9c6.js:33
(anonymous) @ 7851-acd5f0d7727988d3.js:1
(anonymous) @ main-664b20fa39df3103.js:1
(anonymous) @ main-664b20fa39df3103.js:1
n @ main-664b20fa39df3103.js:1
u @ main-664b20fa39df3103.js:1
(anonymous) @ main-664b20fa39df3103.js:1
(anonymous) @ main-664b20fa39df3103.js:1
(anonymous) @ 7851-acd5f0d7727988d3.js:1
(anonymous) @ 7851-acd5f0d7727988d3.js:1
(anonymous) @ main-664b20fa39df3103.js:1
(anonymous) @ main-664b20fa39df3103.js:1
n @ main-664b20fa39df3103.js:1
u @ main-664b20fa39df3103.js:1
Promise.then (async)
n @ main-664b20fa39df3103.js:1
u @ main-664b20fa39df3103.js:1
(anonymous) @ main-664b20fa39df3103.js:1
(anonymous) @ main-664b20fa39df3103.js:1
(anonymous) @ 7851-acd5f0d7727988d3.js:1
(anonymous) @ 7851-acd5f0d7727988d3.js:1
(anonymous) @ 7851-acd5f0d7727988d3.js:1
(anonymous) @ 1230-04040ad560463a89.js:1
eU @ framework-10a404587b40544b.js:9
eH @ framework-10a404587b40544b.js:9
(anonymous) @ framework-10a404587b40544b.js:9
re @ framework-10a404587b40544b.js:9
rn @ framework-10a404587b40544b.js:9
(anonymous) @ framework-10a404587b40544b.js:9
oP @ framework-10a404587b40544b.js:9
eF @ framework-10a404587b40544b.js:9
ro @ framework-10a404587b40544b.js:9
nU @ framework-10a404587b40544b.js:9
nD @ framework-10a404587b40544b.js:9
Y @ _app-695c0ff86878a9c6.js:2

and:

_app-695c0ff86878a9c6.js:28  FatalServerError: Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
    at _app-695c0ff86878a9c6.js:33:254137
    at main-664b20fa39df3103.js:1:115027
    at Object.next (main-664b20fa39df3103.js:1:115132)
    at n (main-664b20fa39df3103.js:1:104710)
    at u (main-664b20fa39df3103.js:1:104907)
    at main-664b20fa39df3103.js:1:104966
    at new Promise (<anonymous>)
    at main-664b20fa39df3103.js:1:104848
    at onopen (_app-695c0ff86878a9c6.js:33:254151)
    at tb (_app-695c0ff86878a9c6.js:33:239083)

Same here, looks like they are doing something new with their side on code execution and we enjoy the results.

Loads of complaints on the Discord server about this and it’s very random. I can paste some scripts with 500 lines of code without a problem in the prompt or 1 “bad” line of code and I get an error. It’s not even about the code sometimes:

1 line of code with a question:


Same line of code without the question:

I read that the problem is browser specific to Chrome/Edge and you can use Firefox as a workaround for now. Would be nice to get an official response from OpenAI about this though as it seems almost impossible to get in touch with them about this.

1 Like

Yes, i also face the same issue. Doesnt seem to be related with the length of the message but something about code. Not sure what…but some code syntaxs seems to break. Also, like others mention, using firefox seems like the only workaround.
This is just one of the issues… Even when in works in firefox, i have observered severe degradation in quality of answers, especially with respect to code. You can ask it for a piece of code, it responds, then you ask for improvements and give it pointers and it doesnt seem to know the previous code it shared.
I repeat again, this has nothing to do with the length, tokens, etc…there is something seriously wrong here.

I get the same error on Firefox. And yes, this is definitely related to Aug 3rd update BUT I only get this when I post more than 100 words. I can literally cut lines of code until it’s less than 100 words, and then it works… This seems like it is related to length of code. Or maybe longer code just has a higher likelihood to trigger this bug.

Regardless, this combined with constant “improvement” of the model is why I unsubscribed.

I haven’t noticed a degrade in quality since the AUG 3rd update, apart from this bug, but overall I’m not happy with how these updates are rolled out. The so called “improvements” of this last update are not “improvements” at all:

  1. Prompt examples. - Useless
  2. Suggested replies - Useless
  3. GPT-4 by default - Useless
  4. Upload multiple files - could be useful if the quality remains stable
  5. Stay logged in - useless
  6. Keyboard shortcuts - could be useful but didn’t miss it

So,

  1. they introduced a bunch of useless features and ruined a perfectly good version doing so.
  2. Introduced a Priority 1 Severity 2 level bug and don’t respond to any messages on the forum, discord or twitter.
    3.Provide no option to switch between versions.

Why is there no community manager, who tests these updates and what on earth is their product owner thinking?

For a SAAS platform like this, they should add the new update as a Beta first for people to try, with an option to switch back to the previous stable version if they’re not happy with it. Then monitor the feedback about the proposed Beta version and tweak it until it’s fit for release, while keeping the community updated on the progress. Better yet, allow people to choose the version they like best and see which is favorite and why. Kind of an A/B community driven testing.

Same here, No HTML works anymore and I keep getting “Something went wrong. If this issue persists please contact us through our help center at help.openai.com.”

I have the same results with your code. You can split it in two for a temporary solution

https://chat.openai.com/share/34c7992a-a60c-4390-95e9-c526555c2ee0

@aolive777

Can you share your HTML code prompt?

You can also avoid this error by throwing some noise into your prompt as seen here:

EDIT I accidentally cleared all of my conversations, re-did it.

https://chat.openai.com/share/edb9ec3f-9496-4047-8a2c-8d16578e79c7

Thanks for reviewing but splitting code up isn’t really a viable solution for coding prompts. I’m using Firefox browser for the time being, which is ok as a workaround. What baffles me is a lack of planning and communication from openai.

Why roll out an update that doesn’t fix anything and then breaks a core function on top of it? No a/b testing, no rollback and no communication. This is not how you run a SAAS business.

I just updated the post with another workaround, just add some noise.

1 Like

I was wrong about not seeing a degrade in quality since the last update. It’s like I’m talking to a mentally crippled toddler again. It can’t even seem to remember the last response or what the topic is. The chat below is about powershell cmdlets and after it keeps giving wrong parameters I provide a list with a few valid cmdlets and parameters. It then responds:

It seems like you’re showing me a list of cmdlets related to PnP.PowerShell 2.2.0, which is part of the SharePoint Patterns and Practices (PnP) library. This library is used for automating tasks and manipulating SharePoint sites, lists, libraries, and more.

However, your message got cut off, and I’m not sure exactly what you would like me to do with this information. Could you please provide more context or clarify what you need? Whether it’s code or an explanation of specific cmdlets or functionalities, I’ll do my best to assist you!


we were trying to get a list all the subsites and it didn’t work because you were using incorrect cmdlets or parameters

And out of nowhere it starts babbling about Wordpress mulitisites…:

I apologize for the mistake in my previous response. Since we are dealing with WordPress multisite, you’ll want to list all subsites using the appropriate functions and tools.

Here’s an example of how you might list all subsites using PHP in a WordPress multisite installation:

function list_all_subsites() {
    $sites = get_sites();
    echo '<ul>';
    foreach ( $sites as $site ) {
        $blog_details = get_blog_details( $site->blog_id );
        echo '<li><a href="' . $blog_details->siteurl . '">' . $blog_details->blogname . '</a></li>';
    }
    echo '</ul>';
}

add_shortcode( 'list_subsites', 'list_all_subsites' );

You can place this code in your theme’s functions.php file or within a custom plugin, and use the [list_subsites] shortcode in your page or post where you want to list all subsites.

1 Like

Yes, I also experience the same. Hope OpenAi realizes this and fixes it soon. Else the plus subscription is simply a waste! :frowning:
I also agree with you, the latest updates literally brings nothing good to the plate. It wasnt worth all the issues its now introduced.

1 Like