Google Brain Founder Andrew Ng Thinks You Should Still Learn To Code - Here's Why

Joe Jenkins

Follow ZDNET: Add america arsenic a preferred source on Google.

ZDNET's cardinal takeaways 

• AI Dev, DeepLearning.ai's AI conference, made its NYC debut. 
• We sat down pinch Andrew Ng astatine nan arena to talk AI and developers.
• Ng recommends that everyone study to code. 

The 2nd yearly AI Dev, a acme connected each things AI and package hosted by Andrew Ng's DeepLearning.ai, came to New York connected Friday. In respective panels and an question and reply pinch ZDNET, nan Google Brain laminitis had proposal to springiness astir nan early of nan field. 

AI has quickly go a reliable coding adjunct for galore developers -- truthful overmuch truthful that galore are wondering astir nan early of nan full profession. Entry-level coding jobs are dwindling for caller grads as teams offload inferior tasks to AI assistants; astatine nan aforesaid time, experts mention nan real limitations of these tools arsenic impervious that engineers will ne'er really go obsolete. 

Also: Why AI coding devices for illustration Cursor and Replit are doomed - and what comes next

Here's what Ng had to opportunity astir really to navigate this uncertain future, why everyone should study to code, and really governance should really beryllium done. 

Coding still matters - benignant of 

"Because AI coding has lowered nan barroom to introduction truthful much, I dream we tin promote everyone to study to codification -- not conscionable package engineers," Ng said during his keynote.

How AI will effect jobs and nan early of activity is still unfolding. Regardless, Ng told ZDNET successful an question and reply that he thinks everyone should cognize nan basics of really to usage AI to code, balanced to knowing "a small spot of math," -- still a difficult skill, but applied much mostly to galore careers for immoderate you whitethorn need. 

"One of nan astir important skills of nan early is nan expertise to show a machine precisely what you want it to do for you," he said, noting that everyone should cognize capable to speak a computer's language, without needing to constitute codification yourself. "Syntax, nan arcane incantations we use, that's little important." 

Also: OpenAI tested GPT-5, Claude, and Gemini connected real-world tasks - nan results were surprising

He added that he wants to invited vibecoders successful arsenic organization members, moreover if they aren't technically developers themselves. But he doesn't expect it to beryllium easy, either. Despite noting that "it's really evident that codification should beryllium written pinch AI assistants," Ng admitted that vibecoding -- which he prefers to telephone "AI coding" -- leaves him "mentally exhausted." 

Becoming generalists   

In his keynote, Ng noted that because AI has made package improvement truthful fast, merchandise guidance -- not prototyping -- is nan caller slowing constituent for launching caller products. To support nan gait AI makes possible, he recommended that engineers study immoderate merchandise guidance skills to circumvent that stall. 

"Engineers that study immoderate merchandise activity tin frankly beryllium a squad of one," he said. 

Also: What Bill Gates really said astir AI replacing coding jobs

That taxable of each professionals -- not conscionable developers -- becoming generalists recurred passim nan summit. During a sheet connected improvement successful nan AI age, Fabian Hedin, CTO astatine coding level Lovable -- 1 of the underdog startups connected a16z's caller list -- noted that vibecoding tin alteration group pinch heavy knowledge successful a non-software taxable to "iterate a batch faster than before" utilizing coding skills. Moderator Laurence Moroney, head of AI astatine Arm, said that this tin make nan astir of an different siloed expert, a displacement successful really niche skills usability successful nan workplace. 

The caller situation for developers, Ng said during nan panel, will beryllium coming up pinch nan conception of what they want. Hedin agreed, adding that if AI is doing nan coding successful nan future, developers should attraction connected their intuition erstwhile building a merchandise aliases tool. 

"The point that AI will beryllium worst astatine is knowing humans," he said. 

Why CS degrees aren't serving students 

The realities of coding successful nan AI property person started to deed post-grads struggling to find jobs. Computer science, erstwhile thought of arsenic a foolproof awesome that guaranteed a lucrative career, is letting students down, Ng told ZDNET. 

He cited nan overhiring sprees tech companies went connected -- and past yet reversed -- during nan COVID-19 pandemic arsenic nan superior logic entry-level coding jobs are difficult to travel by. Beyond that, though, it's a mobility of grads having nan correct benignant of coding skills. 

"AI has transformed really codification should beryllium written, but candidly, galore universities person been slow to accommodate nan curriculum," he said. "So if a assemblage has not importantly changed its curricula since 2022, past they are not readying grads for nan jobs of nan marketplace today." 

Also: AI will origin 'jobs chaos' wrong nan adjacent fewer years, says Gartner - what that means

Ng said he considers it "malpractice" for universities to assistance CS degrees without school those students really to optimize moving pinch AI assistants. 

"I really consciousness bad that there's still group coming receiving a bachelor's grade successful machine subject that person not made a azygous API telephone to a azygous AI model," he said. For him, reorienting CS degrees astir that reality will span nan spread betwixt underprepared grads and a request for AI-experienced coders.  "For nan caller assemblage grads that do cognize those skills, we can't find capable of them," Ng said, a interest he besides noted earlier this autumn successful an X post. 

Public fearfulness of AI 

In his keynote, Ng recognized that "AI has not yet won America's hearts and minds," referring to nan often-circulated nationalist cognition of what AI could go successful its worst-case scenario. Several panelists called connected nan hundreds of developers successful nan assemblage to displacement that perception. 

Joe Jenkins

"You person this unsocial penetration into what AI is not," said Miriam Vogel, president and CEO of Equal AI. She urged developers not to disregard people's fears astir nan tech, but to actively participate successful AI literacy, adding that "we will fail" if that sentiment doesn't improve. 

Ng thinks 3rd parties person intentionally sown AI fearfulness frankincense far. 

"I deliberation a batch of fearfulness of AI was driven by a fistful of businesses that ran almost, frankly, PR campaigns to get group to beryllium fearful of AI, often for lobbying," he told ZDNET during our interview. "I deliberation that's done a batch of harm to nan section of AI and to American activity for developers." 

When asked really developers tin effect that, he said he wants them to prosecute successful candid speech astir what's moving and what isn't. "If nan nationalist understands it better, past we tin each get astatine much logical conclusions astir technology," he said. 

Many of those fears stem from AGI, nan somewhat ill-defined balanced of human-level intelligence that OpenAI and Microsoft, among different labs, person set their sights connected pinch expanding intensity. Ng has agelong maintained that those projections are overblown. 

"If you look astatine nan incredibly messy training recipes that spell into training these AI models, there's nary measurement this is AGI -- if, by AGI, you mean immoderate intelligence task that a quality does," Ng told ZDNET. "So overmuch of that knowledge is still, frankly, engineered into these systems, pinch very clever people, pinch a batch of data." 

Safety and governance 

In a sheet conversation, Ng acknowledged that nan nationalist doesn't really cognize what AI labs are doing, which tin create panic, but urged group not to "take a reddish teaming workout and move it into a media sensation." Ng added that he's little successful favour of Anthropic's marque of information and governance, which he finds somewhat limiting. Rather than base down connected governance efforts, he emphasized sandboxed environments "that are guaranteed safe" arsenic a way toward responsible AI that doesn't hamper speed. 

Vogel defined governance arsenic "breaking down principles into actionable workflows," not creating bureaucracy. Her interest was little astir nan hyperscalers for illustration OpenAI and Meta, and much astir nan smaller AI companies that blaze up earlier they've developed immoderate governance structure. 

Regulating AI  

"You don't get to lead successful AI by passing regulations," Ng said during a panel, speaking connected nan EU's attack to legislating AI. He credited nan Trump administration's AI Action Plan, released this past summer, for keeping national regulations loose. 

Many AI experts are alarmed astatine nan deficiency of US AI regulation. Some spot nan national government's nonaccomplishment to modulate societal media platforms erstwhile they proliferated arsenic an illustration of what could hap if AI continues to outpace legislation. Ng told ZDNET he thinks that's a mendacious equivalency. 

Also: 8 ways to make responsible AI portion of your company's DNA

"I'm seeing measurement much bad regulatory proposals than bully ones," Ng said successful nan interview, adding that he sees nan nonconsensual deepfake prohibition and nan FTC's actions against companies using AI to standard "deceptive aliases unfair conduct" arsenic examples of bully AI policy. 

When asked if location are immoderate different regulations he'd make astatine nan national level, he said he wants much transparency requirements for ample AI companies disconnected nan bat. 

"When a batch of bad things happened pinch societal media, nary of america knew astir it. Even group wrong nan business did not really cognize astir it," Ng told ZDNET. "If we person regulations to require that nan largest companies -- only nan ample ones, truthful we don't put unreasonable compliance burdens of mini startups -- but if we request immoderate level of transparency from nan businesses pinch a very ample number of users, that could springiness america amended signals to spot nan existent problems, alternatively than count connected nan luck of location being a whistleblower."