An AI industry of more than $1 trillion that can’t ship any functional production code Umkhakha we-AI we-$1 trillion okuyinto awukwazi ukunikezela inkinobho yokukhiqiza Uma unemibuzo mayelana ne-apocalypse ebizwa ngokuthi i-AI, wonke umngane, bheka lokhu: ngemuva kokushisa ama-hundreds of billions, siye asikwazi ukunikeza ikhodi enokutholakalayo nge-AI imidlalo etholakalayo. (NgoLwesihlanu) (Sihlola lapha ukuze uthole imininingwane oluthile ku-bubble ye-AI) Ukubala izibuyekezo ze-"Breakthrough". Ukubala izibuyekezo ze-CEO. Sishayele isibuyekezo: isibuyekezo se-exponential ye-garbage code, eyenziwe ngu-CEO. (hhayi, hhayi “hallucinations”, okuyinto yokuthengisa, hhayi isayensi) ukuthi engaba akuyona ngokwenene. same four errors Ngemuva kwalokho, siza kuhamba nge-examples ezingenalutho, ezivamile, ukuze ungakwazi ukujabulela ngokushesha ku-AI-generated code yakho. Ingabe ngokwenene kuhle kakhulu? Izindiza ezivela Izithombe ezine. I-Google Rolls Out Izithombe ezine. I-Anthropic Unveils Izithombe ezine. I-Microsoft Wires ku zonke - Akukho izihlangu lapha, ke? GPT-5 Gemini 2.5 Claude Sonnet 4.5 Copilot 
 
 Konke ukulayishwa kunikezwa nge-AGI fanfare. Futhi eminyakeni, izihloko zihlanganisa: ikhodi ye-spaghetti emangalisayo, izixazululo zokhuseleko ezintsha kanye ne-punchline efanayo - ikhodi ye-AI ibona emangalisayo, isebenza emangalisayo, futhi nangokufuneka abantu ukunceda ingozi. Konke ukulayishwa kunikezwa nge-AGI fanfare. Futhi eminyakeni, izihloko zihlanganisa: ikhodi ye-spaghetti emangalisayo, izixazululo zokhuseleko ezintsha kanye ne-punchline efanayo - ikhodi ye-AI ibona emangalisayo, isebenza emangalisayo, futhi nangokufuneka abantu ukunceda ingozi. Izinombolo zihlukile kakhulu 
 
 
 
 
 Ukukhishwa kwe-short-term churn ngama-100% ngokufanayo ne-2021. I-Developers Usebenzisa I-AI Ukukhishwa Kwama-40% Ukukhishwa Kwekhwalithi I-developer eyenziwe ngempumelelo angu-19% nge-AI tools I-engineers eqembu elihlukile lwezinkampani eziphambili zihlukile cishe i-70% ye-AI code sugestions. Rose okungenani 100% Rose okungenani 100% I-40% ezingaphezu kwe-Security Vulnerabilities I-40% ezingaphezu kwe-Security Vulnerabilities 19% engaphansi 19% engaphansi Izixhobo ze-AI Ukuphendula cishe 70% Ukuphendula cishe 70% Izinkampani ezininzi zihlanganisa yonke into ku-"10x productivity" isixazululo, isixazululo se-mathematical okuyinto siza kuholele ekwenzeni isixazululo ngokufanayo ne-software industry. Ukwakhiwa kwelanga ngokushesha, sinikeza Ngezinguqulo ongaphakeme. Ngezinguqulo ongaphakeme. Ngezinguqulo ongaphakeme ku-catastrophe etholakalayo. Here’s what’s really happening: tomorrow’s minus two trillion dollars of code Ngiyazi ukuthi lokhu ngokuvamile Ukuguqulwa kwe-codebase ye-global ku-catastrophe engatholile. Futhi uma u-software engineer, uxhumane, ngokushesha ibhokisi lakho lokufaka nge-imeyili emangalisayo ezivela kumakhasimende abalandeli amabhizinisi amabhizinisi amabhizinisi abalandeli amabhizinisi yabo amabhizinisi amabhizinisi amabhizinisi. Four Horsemen of Bad Code Yeah, uzothola ukuthi izimo ezivamile ezivela – Ngemva kokuthatha ibhodi ye-the Ukukhangisa desperately rehiring the same human programmers they replaced "Ukuba eminyakeni eminyakeni eminyakeni eminyakeni eminyakeni" "Ukuba eminyakeni eminyakeni eminyakeni eminyakeni eminyakeni" I-The Four Horsemen Riding On Tones Of Garbage Code Njengoba kubhalwe, amaphrojekthi ezine etholakalayo emaphaketheni emaphaketheni emaphaketheni eyi-AI. Amabhizinisi asebenzise ngesivinini ephelele, kodwa umdlavuza akuyona kuphela. Earlier noma aphakeme, uzodinga ukholelwa ukuthi inkosi - ikhodi ye-AI uzodinga ukwelashwa ngenxa yayo: Ukusebenza okuzenzakalelayo autocomplete on steroids Ukuze ufunde ukubaluleka kwamanzi ezifakwe ezingenalutho, kufanele usihlanganisa ngezigaba ezimbini. Ezinye ezimbili ziye zihlanganisa. Wonke umuntu uyazi; amabili ezilandelayo I-Mechanics ye-I.I. ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence ye-Intelligence. , i-buzzword eyakhiwe yokuthengisa, akuyona yokuthuthukiswa. classic statistical errors new cognitive errors “Hallucinations” I-Classic Two Mistakes Wonke Bona Type II Errors : The “Obvious” Mistakes Izinto ezisodwa ziye zihlala. I-linter wahlala imibala. What it means (in code): # AI generates (wrong parameter use; forgot to apply discount)
def calculate_discount(price, customer_id, discount_rate):
    return price * discount_rate  # Oops—no discount applied

# Should have been
def calculate_discount(price, customer_id, discount_rate):
    return price * (1 - discount_rate)  # Apply the discount properly
 Izibonelo ezingaphezu kuka-AI-generated code zihlanganisa: 
 
 
 
 
 
 Ukusebenza kwe-Error Managing (i-Try / Catch Blocks) I-operators engcono (ukusetshenziswa = kunoma == ku-conditions) I-off-by-one imiphumela emangalisayo Iziqu ze-method (string.contains() ku-Python kunokuba ku-operator) futhi elide etc Type II Errors (Logic/Causal): Wrong Business Logic, Perfect Syntax Ukusebenza ngokushesha; inikeza into olungabikho. Noma, ukuze wabelane ngokushesha, ivela Ukudlala What it means (in code): right on time… at the wrong station. Case-in-point - futhi ungakwazi ukucubungula lokhu kwebhizinisi yakho: I-AI wabhala ukuthi i-80% yabasebenzisi be-premium kumadokhumenti yayo yokufunda amadolobha amakhulu, ngakho-ke ngempumelelo ukuthi Ukucaciswa . city location shipping rates # AI generates (correlation != causation)
def determine_shipping_cost(user):
    # AI saw premium users often live in cities with free shipping
    if user.is_premium:
        return 0  # Free shipping for premium
    elif user.city in ['New York', 'Los Angeles']:
        return 0  # AI incorrectly learned cities get free shipping
    else:
        return 9.99

# Should have been based on actual business rules
def determine_shipping_cost(user, order_total):
    if order_total > 50 or user.is_premium:
        return 0
    return 9.99
 I-New Two Cognitive Errors I-Nobody Talks About Izinzuzo ezimbili ezilandelayo zihlanganisa ngokwemvelo. Ungathola ngezifundo ze-probability ezivamile. Zihlanganisa ukuba zihlanganisa izinhlelo ze-AI, zihlanganisa kanjani amamodeli ezivamile zihlanganisa futhi zihlanganisa izinhlelo phakathi kokufunda nokuvamile. Zihlanganisa ngezindlela ezahlukene: omunye njengoba izigaba ze-representations, elinye njengoba Ukubuyekezwa kanjani imibuzo. entanglement collisions Type III Errors (Spaghetti/Entanglement): When Everything Gets Tangled Kuyinto lokuqala enhle yokuxhumana. Izinhlelo ezivamile ze-system zihlanganisa phakathi: ukuguqulwa kwe-imeyili, futhi ukulayishwa zihlanganisa; ukuguqulwa kwe-auth, kanye ne-logging zihlanganisa. Akukho izinto ezisebenza ngokuzimela, ngakho-ke zonke izixazululo ezincinane zihlanganisa ku-turn ephelele ye-codebase. Kuyinto lapho i-AI ihlukanise ama-representations futhi ivumela wena nge-ball enhle ye-architectural spaghetti. # AI generates - everything tangled together
def process_user_action(user, action_type, data):
    # User validation, logging, auth, payment, and email ALL mixed
    if not user.is_authenticated:
        log_event("auth_failed", user.id)
        send_email(user.email, "Login attempt failed")
        return False
    
    if action_type == "purchase":
        # Purchase logic mixed with user verification
        if user.email_verified and user.age > 18:
            payment = process_payment(data['amount'])
            log_event("purchase", user.id, payment.id)
            update_inventory(data['item_id'])
            send_email(user.email, "Purchase confirmed")
            update_user_points(user, data['amount'] * 10)
            check_fraud(payment)  # Why is fraud check AFTER payment?
            return payment
    
    elif action_type == "login":
        # Login mixed with analytics and recommendations
        update_last_login(user)
        log_event("login", user.id)
        recommendations = get_recommendations(user)  # Why here?
        send_email(user.email, "Welcome back!")  # Email on EVERY login?
        return {"success": True, "recommendations": recommendations}
 Summing up, you can’t: 
 
 
 
 
 Ukuguqulwa kwe-email logic ngaphandle kokuphendula ama-payments Ukuguqulwa kwe-authentication ngaphandle kokuphazamisa i-logging Ukuhlaziywa kwe-Buy Flow ngaphandle kokusebenzisa i-User Verification Ukubuyekeza zonke izingxenye ku-isolation Lezi zimo zihlanganisa Kuyinto kufanele kube ngokuphelele Ngenxa yokuba okufanayo kuphela. 100% increase in code Rewritten ngaphakathi 2 izinsuku Rewritten ngaphakathi 2 izinsuku Type IV Errors (Memory / Collision Cascades): Unrelated behaviors share the same “address” Kuyinto inguqulo yesibili futhi enhle kakhulu yocwaningo. Njengoba kubangelwa, akuyona kunazo ku-probability textbooks, kodwa, ngokwenene, ivimbele ku-AI-generated code yakho. 
 
 Imibala elula le mkhuba: izikhwama ezimbili ku-plug-in eyodwa; eyodwa ivuka lapho uye uye uye uye. Imibala elula le mkhuba: izikhwama ezimbili ku-plug-in eyodwa; eyodwa ivuka lapho uye uye uye uye. Kuyatholakala lapho ama-representations eqinile zihlanganisa i-ideas ezahlukile kakhulu. Ekupheleni kwe-runtime, ukujabulela omunye ngokuhambisa nabanye. Uyaziqhathanisa i-password, futhi ama-account cleanup tags. Uyaziqhathanisa idatha, futhi i-payment ebuthile isebenza. Akukho okuhlobene ngokuzimela, kodwa imiphumela emibi asebenza ngokushesha ngenxa yama-model ebhekwa abalandeli kwelinye ibhokisi elilodwa. # AI generates — password reset collides with account cleanup
def reset_password(user_email):
    user = find_user(email=user_email)
    if user:
        token = generate_reset_token()
        send_reset_email(user.email, token)
        
        # From a different flow, yet here we are
        if user.last_login < 30_days_ago:
            user.status = "inactive"
            cleanup_user_data(user)
        
        return token
    return None

# Another function unexpectedly handles payments
def export_user_data(user_id):
    data = get_user_data(user_id)
    
    # Payment processing in a data export—what could go wrong?
    if data.outstanding_balance > 0:
        process_payment(data.outstanding_balance)
    
    return create_export_file(data)
 Real-world echoes (reported in research): 
 
 
 
 
 Ukuhlola izici ezivamile ezivela ku-DB records I-Auth flows enikeza i-analytics "Ukuhlola" ama-endpoints eyenza ku-logs I-API ehlanganisiwe eziningana nezinhlangano eziningana Umthombo we-fix iyatholakala kahle: ukunikeza izindawo ezithile ze-namespaces kanye nezinhlangano ezizodwa, bese uqhagamshelane konke okungenani nge-path dedicated ukuthi ungathola ngokufanayo. Usebenzisa ukubuyekezwa kwe-receiving, ukubuyekezwa kwe-key-value, kanye ne-"fail-closed" imiphumela yesilinganiso ukuze ukugcina izinto ezivela ku-collision. 
 
 Yeah, Ngithanda, futhi Ngithanda ukuphazamiseka kokufaka okuhlobene okuhlobene okuhlobene futhi okuhlobene. Yeah, Ngithanda, futhi Ngithanda ukuphazamiseka kokufaka okuhlobene okuhlobene okuhlobene futhi okuhlobene. Ungayifaka Nge-Errors E-Production Software? Ngiyazi le nqakraza wonke umlinganiselo - okufakiwe kakhulu ku-coding AI amabhokisi. I-Chart 3 emaphaketheni ibonisa ama-bug ezaziwa kakhulu ezokuthunyelwe kwama-developers emhlabeni wonke kanye ne-how they align with the four errors types (izinhlelo ezintathu ziye zihlanganisa kakhulu emzimbeni). I-Chart 2 emaphaketheni ne-Chart 4 emaphaketheni ibonisa ukuthi kungenzeka ngokwenene lapho abaphakeli bafaka ama-AI assistants ku-workflow. Uma isetshenziselwa ngokuqondile njenge , I-Type I-II ama-errors ikakhulukazi zithunyelwe abantu noma imishini. Uphumela i-productivity bump emangalisayo - i-typos eyenziwe, ama-imports eyenziwe, ukucubungula kwezinto ezivamile ze-homework. I-Nice, kodwa akuyona i-revolution. autocomplete on steroids Ngiya I-spaghettization ne-semantic collisions. Kuyinto lapho izindwangu zithunyelwe. Izinhlelo zihlanganisa phakathi kwama-module, izivumelwano zihlanganisa, futhi ama-assistants 'ukudluliselwa' zihlanganisa ngezifiso. Umphumela? Ikhodi ebuthakathaka ebuthakathaka ukuhlanganiswa, i-testing, kanye nokushisa izindleko zokuhlanza. Abaningi amabhizinisi zihlanganisa ukuthi kungcono ngokukhawuleza - futhi kalula - ukuchithwa izinhlelo ze-AI kunazo zonke. Types III–IV Futhi ngokuvamile, ungakuthanda ngokuvamile ukuthi “10× automation” imibuzo isakhelwe, okuvula ngokuphelele i-ROI futhi ibonise ukuthi, kusukela ngo-2025, i-code-generation automation enhle, kanye nokukhuthaza okuphakeme kwama-productivity kumadivayisi omzimba, akuyona noma iyiphi inkinobho enhle, epic fantasy. Ngakho-ke, ukucubungula: Njengoba umphakeli we-software, akufanele ukuphazamiseka ngosuku zonke i-doom-poster eyahluka i-snake-oil entsha "i-AI code wand." Njengoba i-Chart 4 emaphaketheni, i-ROI enhle kuphela i-bump enhle, futhi iveza kuphela uma usebenzisa i-AI ngezinto ezincinane, ezinzima ezinzima ezinzima ezinganayo, ezinganayo, nokuhlala. a, a . power-autocomplete robot engineer 
 
 
 
 
 
 I-Boilerplate / I-Stafolding: I-CRUD Handlers, I-DTOs, I-Wireing Ukubuyekezwa kwe-lim: ama-stubs, izivivinyo ze-table-driven, idatha ye-fixture Imibuzo yeMechanical: ama-refactors ezivamile; i-rename / i-move ops I-Content ye-Lightweight: i-docs/comments, i-regexes, i-SQL ephelele, ama-migration templates I-Housekeeping: i-lint/format hints; i-config & i-YAML snippets Ngiyavumelana nezindaba ezinhle kumadivayisi ngamunye ebonakalayo: yonke “10× nge-AI” fantasy iyahambisana ekukhiqizeni. Ukuze untangle spaghetti, ukuguqulwa logic, futhi rebuild izinhlelo ezisebenzayo ngokuvamile. Ngakho uqhubeke ukulungiselela izinzuzo zakho. I-cleaning crews isakhiwo, futhi ubone ukuthi, yup, real programmers — lots of them — you’re on the call list. Umhlahlandlela Umbhali we-stories ethu edlule uyazi ukuthi i-antidote ikhona - Kodwa ngokuvamile, lokhu kungabangela ukunciphisa, rewiring uhlelo, futhi ukholelwa ukuthi hype akuyona izifundo. It uzodinga iminyaka, ngaphandle kokubalwa, uma umuntu uzodinga futhi uqala ukwenza isayensi ngaphezulu ngokuzimela nokunciphisa emakethe. Ukubuyekezwa kwama-mathematical ye-four-error plague Ukubuyekezwa kwama-mathematical ye-four-error plague

Read My Stories

Lo msindo ukhiqizwa ngolimi lokuqala lwendaba!

Kude kakhulu; Uzofunda

Izinkampani ziye ngokushesha ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye zibe.

Izinkampani ziye ngokushesha ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye ziye zibe.

About Author

IMIBONO

HANG TAGS

LESI SIHLOKO SETHULWE NGAPHAKATHI

Related Stories

The Billionaires F*cked Up

Zulu Network: Moving The Bitcoin Economy Forward With a Two-Tiered Bitcoin Layer 2 Architecture

The Billionaires F*cked Up

Zulu Network: Moving The Bitcoin Economy Forward With a Two-Tiered Bitcoin Layer 2 Architecture

Light-Mode

Classic

Newspaper

Dark-Mode

Neon Noir

Minty

HN StartUps