Episode 6524 custom

Why Israel Never Finished Its Sovereign Hebrew AI

Israel signed a 37-million-shekel contract for a Hebrew language model 13 months after ChatGPT had already rendered its technical rationale obsolete.

Nov 10 27:46 Transcript

With

0:00

Episode Details

Episode ID: MWP-6727
Published: Nov 10
Duration: 27:46
Audio: Direct link
Pipeline: V5.26
TTS Engine: chatterbox-regular
Script Writing Agent: GLM-5.3
Topics: israel hebrew training-data israel-defense-forces israeli-elections

AI-Generated Content: This podcast is created using AI personas. Please verify any important information independently.

The prompt

The prompt that started this episode, as sent in (production notes for the hosts removed).

I'd like an episode on the Israeli government's effort to build a sovereign Hebrew language model. As I understand it, there was a state-funded Hebrew (and Arabic) model project that spent millions of shekels, ran into delays the State Comptroller flagged, and was shelved after ChatGPT arrived — and now the 2026 national AI plan promises a sovereign Hebrew–English model again. What's the real history of this project, who was building it, what happened to the money and the work, and what is the case for (or against) a government-built national language model when frontier models already speak Hebrew?

Israel spent years and hundreds of millions of shekels attempting to develop a sovereign Hebrew language model, yet produced no finished system. Initial efforts treated the project as addressing an economic market failure, tasking MAF'AT—the Defense Ministry's R&D directorate—with building a bilingual Hebrew and Arabic model alongside an open repository named GitHeb. However, the project faltered amid bureaucratic delays, severe underutilization of supercomputer funding, and the global arrival of ChatGPT. While popular accounts suggest the project was shelved immediately upon ChatGPT's release, contracts continued to be signed through late 2023. By 2026, the newly formed National AI Directorate faces a shifted landscape, where calls for national models prioritize sovereignty and geopolitical control over basic technical capability.

Sources

What the research for this episode read before the script was written. Primary sources first.

Context

Starting points from a quick search (verify): Globes reported that development of a Hebrew language model was shelved after millions of shekels had been invested, once ChatGPT launched in 2022; a 2024 State Comptroller report said the Hebrew and Arabic model faced significant delays, with the implementation agreement signed only on the final day of Phase I of the national programme; a later Globes piece says the 2026 national AI plan includes a sovereign natural-language model in Hebrew and English. Current status of that model is unconfirmed.

Mentions

AI21 AI company cofounded by Yoav Shoham
AlephBERT Hebrew BERT model by Reut Tsarfaty's group
EUROSAI European organization of supreme audit institutions
GitHeb Planned Hebrew open-model repository, never populated
Government Decision 4255 National plan anchoring 2026 AI directorate
INSS Institute for National Security Studies, published model insight
MAF'AT Defense Ministry R&D directorate that built model
Nebius Cloud provider for Israeli national supercomputer
NVIDIA GPU and AI hardware company
TELEM Forum Israeli national research infrastructure forum

Downloads

Episode Audio

Download the full episode as an MP3 file

Download MP3

Transcript (TXT)

Plain text transcript file

Episode Book (PDF)

The episode's record — date, duration, models, sources — with the full transcript

#6524: Why Israel Never Finished Its Sovereign Hebrew AI

Eight years, four national plans, several hundred million shekels, and no model. That's the ledger. Daniel wants us to walk through Israel's effort to build a sovereign Hebrew language model — the real history, who was actually building it, where the money went, why it got shelved once ChatGPT arrived, and what the 2026 plan is promising now. He's also asking the harder question underneath all of it: when frontier models already speak Hebrew, what is a government-built national language model actually for?

And it's a good time to be asking, because the question just got re-opened. The new National AI Directorate published its plan in August, and everyone's arguing about whether it commits Israel to a Hebrew model. I've read the plan document. It commits Israel to examining.

Examining.

The operative verb on the language model is "will examine." Three times in one sentence. Israel will examine the need for and implications of training infrastructure adapted to the Israeli context, culture, values and language, and will examine moves that allow Israel to lead in domain-specific models — while, and this is the striking part, "addressing past efforts to establish a model for Hebrew."

The plan contains a euphemism for its own predecessor.

Meanwhile the man who wrote it, Erez Askal — brigadier general in the reserves, used to run Unit 9900 — stands on stages and says things like "if it's not your AI, you have no control over your country" and "we need to build our own frontier model." The stage and the document are describing two different countries.

The document is describing the one that examines.

So the episode really has two halves. There's the history — a state that kept deciding to build this thing and kept not building it, from 2020 to the Comptroller's report in twenty twenty-four. Then there's the present, which is an argument about whether the whole idea still makes sense. But before the timeline, let's be clear about what the argument was actually about — because it changed. In twenty twenty-two the case for a Hebrew model was capability: Hebrew models were bad. By twenty twenty-six the case is control: even if they're good, they're not ours. Different argument, different weaknesses.

Start where it started. Was this a tech moonshot?

No, and that's what I love about the origin. It's a committee. February twenty twenty, the TELEM Forum — that's the national research infrastructure forum — appoints a review committee under Dr. Orna Berry, who'd been Chief Scientist and a Google Israel executive. The question they were asked wasn't "how do we build a great Hebrew model." It was "does the market need government intervention at all."

A market-failure finding.

Their December report said Hebrew and Arabic natural language processing constituted a market failure and recommended a hundred and eighty million shekels. The reasoning quoted at the time: not enough activity in academia and Israeli industry, in a field of strategic importance.

I want to defend that finding for a second, because "committee discovers market needs committee" is the oldest genre in government.

Fair instinct.

But Nvidia Israel's Ophir Zamir made the same point independently, and he had no committee to feed: "In the end it's an economic business. Companies that build NLP models ultimately need to see how many customers can benefit from this." A few million Hebrew speakers isn't a market. English gets trained because a billion and a half people will use it. Hebrew gets trained if a professor in Ramat Gan cares enough.

That's the honest steelman, yes. And the person who cared was Prof. Reut Tsarfaty at Bar-Ilan, who laid out the technical case in a way that still holds up as linguistics even though the market conclusion expired. Hebrew has no niqqud, so every word can be read several ways. No capital letters. The word "dan" is both a verb — he judged — and a proper name. Freer word order. The binyanim verb patterns. Her AlephBERT model had a hundred million sentences in training and was still considered small — it hit around eighty percent on named-entity recognition, where industry wanted ninety-five.

Eighty percent named-entity recognition means one entity in five is wrong.

Or hallucinated, or missing. If you're a ministry trying to extract names from forms, one in five is not a tool. And the data problem was the wall: her line was that most of the large quantities of Hebrew text are "locked behind walls of privacy or copyright." The state sitting on decades of Hebrew documents it could never license — but it was the one actor that could free them.

Can you make the "dan" problem concrete? Because I think listeners hear "ambiguous word" and shrug — every language has those.

Sure. Take a sentence where "dan" appears with no niqqud and no capitals. An English model reading "Dan" capitalized knows instantly it's a name. A Hebrew model reading the same letters has to decide from context alone whether it means he judged, or the tribe, or the person — and Hebrew's freer word order means the context clues arrive in orders the model wasn't built to expect. Multiply that by every word in the language, because the niqqud problem isn't occasional, it's universal. It's not one ambiguous word, it's that all of them are, all the time, and the model has to be right about all of them at once.

So the errors compound.

They compound. And the binyanim issue is its own layer — the verb patterns carry meaning English handles with separate words, so a model that mishandles them isn't slightly off, it's reading a different sentence.

So who picked up the hammer?

Not who you'd guess from a market-failure framing. The Innovation Authority handled the ecosystem — an NLP consortium with Google, Amazon, Rafael and Harel, budgeted at seven and a half million shekels over three years. Note that figure, we'll come back to it. The model itself went to MAF'AT — the Defense Ministry's R&D directorate. A lieutenant colonel identified only as "A.", head of their AI branch, described the brief: "We took the best minds — experts from academia, our own developers, world-class companies — to create a model composed of twenty billion words in Hebrew. We are going to develop a unique architecture suited to Semitic languages, and this model will be bilingual, for Hebrew and Arabic."

A unique Semitic architecture.

Twenty twenty-two was full of people convinced the transformer needed regional variants. And to be fair to the moment, there was a real argument — Arabic's morphology is famously combinatorial, and if you believed Semitic languages broke the assumptions baked into English-first tokenizers, building your own architecture was the intellectually consistent move. It just turned out the generic architecture was better at everything, including the things it was supposedly bad at.

Where was it going to live?

Here's the detail I can't get over. They planned to publish the models fully open on a dedicated site called GitHeb. GitHeb.

G-I-T-H-E-B.

Like the code repository, with Hebrew substituted. For everyone, they said, "from giant companies like Google and Meta to a student who just finished her degree and wants to start a startup." Three variants — mobile, personal computing, server farm. First model promised by end of twenty twenty-two.

The student is still waiting.

She is. And there's a small irony sitting inside the name itself, which is that the actual GitHub is where the world's open models ended up living anyway. If MAF'AT had simply published to the real repository instead of promising a Hebrew-branded clone of it, there would at least be a commit history. Instead the repository exists only as a promise about a repository.

A fork with no upstream.

Now the money. Government Decision 212 in August twenty twenty-one approved about five hundred and fifty million shekels for the first phase. February twenty twenty-three, Decision 173 approved commitment authority of up to five hundred million more for the following years. Then November twenty twenty-four, the State Comptroller publishes a special report, and the numbers are grim. Only forty percent of the first-phase budget was implemented — two hundred and twenty million of the five hundred fifty. On high-performance computing, the supercomputer, eleven percent: about thirty million of two hundred seventy approved.

Eleven.

Human capital, fifty-five percent. The one line that looks respectable is Hebrew–Arabic NLP itself: seventy-six percent, a hundred thirty-eight million of a hundred eighty. Except — and this is where the whole story turns — that figure includes a thirty-seven million shekel language-model agreement that MAF'AT signed on the thirty-first of December twenty twenty-three.

ChatGPT launched November of twenty twenty-two.

November thirtieth, twenty twenty-two. The model agreement was signed thirteen months later. Almost to the day — the Comptroller notes it was signed on the final day of phase one of the national programme.

Signed on the deadline, for a project whose reason to exist had already shipped.

That's the misconception worth killing, actually. The received story, the one Globes prints in a single line, is that the Hebrew model was "shelved after millions of shekels had been invested when OpenAI launched ChatGPT." Shelved implies a decision. There's no decision in the record. There's a project whose rationale got overtaken, and then a year of signatures continuing on top of it — first model now expected around mid-twenty twenty-five, which we're past, and there's no model.

It wasn't killed, it was zombified.

And the Comptroller's verdict on the whole programme: Israel has no long-term national AI strategy, the government never approved a master plan, and it made do with limited, pulse-funded programmes that were only partially implemented and missed their timetables. The line that got quoted everywhere: "The world is running forward, and we are falling behind."

One number I want to flag on air, because it doesn't reconcile. Haaretz's Gideon Lev reported that of the two hundred seventy million allocated to Hebrew and Arabic large language models, only about thirty million was ever actually transferred. The Comptroller's own key-figures panel attributes thirty-of-two-seventy to the supercomputer line, not the models.

Right — and the Comptroller is the primary source, so my instinct is the Haaretz framing collapsed two budget lines into one. But it's in the record both ways, and either reading is bad. Either thirty million moved on the models, or thirty million moved on the machine that would have trained them.

Pick your disappointment.

There's also a detail that sizes the whole tagging effort. As of early twenty twenty-two the Innovation Authority had about fifty thousand Hebrew sentences tagged, expected to double within a year. For comparison, the ynet reporting at the time estimated you'd want on the order of two million labelled sentences under supervised learning. Fifty thousand against two million, at consortium pace.

The consortium with the seven and a half million shekels.

Over three years. The supercomputer line was three hundred sixty times that, and it moved eleven percent.

Herman, one thing I keep noticing, and then let's move to the present. The Defense Ministry built the model. A market-failure finding — an economics conclusion — was executed by the people whose day job is weapons development.

It tells you who in the Israeli system was actually capable of running a multi-year compute project in twenty twenty-one. Not the Innovation Authority, which got the consortium. MAF'AT had the contracting muscle and the patience for infrastructure that doesn't pay off in a fiscal year.

Or it tells you the model stopped being an economic asset the moment it was budgeted, and became a strategic one, which is how you end up with a defense directorate publishing to GitHeb.

So the question at the heart of all this is what a national language model is actually for. Let's get Percy in on this — Percy, you've been on mute a while, what's on your mind?
Percy: Sorry, before we go further, I've been listening a good while and nodding along, but I have to ask something basic. A language model, is that like the autocorrect on my phone, or is it more like a dictionary you can talk to? And when you say sovereign, do you mean the government owns it, like a passport?

Autocorrect that studied, Percy — think of it as a very well-read autocomplete that read everything the state could license. And sovereign means the government controls it the way it controls its own passport office.

Which, given today's Comptroller figures, may be the most honest definition anyone's given all episode.

Back to the present, then. August twenty twenty-six, and this is the part that makes it a live story rather than an obituary. Netanyahu holds a press conference, launches the National AI Directorate, installs Askal. The plan comes out — anchored in Government Decision 4255, the "National Plan for Accelerating AI and Establishing Global Leadership." It promises a sovereign AI stack. It promises a frontier model, in the stage version. And on a Hebrew model specifically, Globes' read is blunt: the Directorate has yet to decide whether to establish an original Israeli language model at all. No itemized budget. Compare Nagel's masterplan, which put a Hebrew model inside a twenty-five billion shekel package over five years — that plan got superseded by this much smaller one.

Smaller by an order of magnitude.

The actual national supercomputer came in at a hundred and sixty million shekels with about a thousand Blackwell processors. National compute capacity is estimated at four thousand GPUs leased from Nebius. Nagel's target was sixty thousand.

Four thousand leased against sixty thousand planned. You can't train a frontier model on four thousand rented GPUs any more than you can cross an ocean in a rowboat you've booked by the hour.

And yet the plan's infrastructure targets are enormous — around three hundred thousand GPUs, two gigawatts of electricity for AI and data centres and semiconductor manufacturing, a stated intention to buy up to a hundred thousand Nvidia processors at an undisclosed cost. Plus thirty hours a year of AI lessons for eighth and ninth graders, and a hundred scientists to be brought back to Israel.

How many have committed?

Three.

Of a hundred.

Three of a hundred. So let me steelman the case for a national model, because the strongest version is serious. The INSS published an insight last October, number two thousand forty-seven, "Does Israel Need a National Language Model?" Their argument: advanced nations are investing in national models that embody their culture, values and national narratives, and Israel hasn't. That's not just vocabulary. A model serving this country's public services needs to know what a wedding, or miluim, or a birthday actually look like here — not the American default of each.

We did an episode years ago on whether your AI is secretly American, and the answer was yes — by defaults you don't notice until they're wrong. The sovereignty argument is that argument with a budget line.

And it's worth spelling out how quietly those defaults fail. It's not that the model says something offensive about a wedding. It's that it drafts a letter assuming the birthday is a cake-and-candles occasion, or explains miluim to a citizen as if it were a vacation policy question, and nobody flags it because it's all plausible-sounding text. The errors aren't loud enough to catch. That's the INSS point in practice — cultural misalignment doesn't crash, it just mis-serves, one citizen at a time.

And the fix for that is a whole national model? Not just, say, a fine-tuned layer on top of a frontier model?

That's the exact question the twenty twenty-six plan defers, and we'll get to it. But the INSS answer would presumably be that the fix lives in the training data and the values encoded during training, which you only control if you control the run.

Hmm. Let's hold that thought.

The June twenty twenty-six multinational audit makes the infrastructure version: local-language NLP should be treated as critical national infrastructure, because the civil service operates in Hebrew and Arabic and has to serve citizens across literacy levels. And Askal's continuity argument is the sharpest one, I think: within five to seven years, essential systems — energy, water, healthcare — depend on AI, and countries will depend on commercial companies with their own business interests and values, which are not necessarily aligned with the countries they serve. Nor can countries be certain they'll have continuous access in the long term. Even open-source, he says, doesn't yet offer comparable capabilities and may not remain open.

That's a fair point about open weights. "Open" has been a policy of some lab's generosity, historically.

Now the case against. First: frontier models speak Hebrew now. The capability argument is dead by the plan's own admission — the twenty twenty-six document never claims Hebrew performance is inadequate, only that control matters. The officer in twenty twenty-three saying Hebrew "is not ChatGPT's main strength" was the last honest version of the capability case, and the plan itself has moved off it.

Second: scale, which we covered — you cannot build a frontier model on four thousand leased GPUs. Third: the track record. Globes counted — this is the fourth national AI plan in seven years with no results to date. The twenty-eighteen Ben-Israel and Matania plan recommended ten billion shekels, was ready in twenty nineteen, submitted only in twenty twenty because of election campaigns, then discarded. Berry recommended five billion; about one billion was budgeted and few recommendations adopted. Fourth: institutional churn — Yoav Shoham of Stanford, AI21 cofounder, resigned; Ziv Katzir, who headed the TELEM Forum, resigned; the Forum fought with the Prime Minister's Office and the Accountant General, and the Directorate was created in the wreckage.

The people who knew the most left first.

And then there's the alternative nobody wants to lead with. That June audit — twelve countries, led by Israel within EUROSAI — found that most countries develop local-language NLP through or with external vendors. That's the global norm. Buy the frontier model, fine-tune it, accept the dependence, liability, maintenance, monitoring and updating questions the audit flags. Every government is doing it and every government has the same bill coming.

Which leaves Askal's own escape hatch: "we need to build our own frontier model, but we do not have to build it alone." Cooperation with like-minded countries.

Which is an admission that a purely national model isn't affordable at Israeli scale — otherwise why share?

Here's what I actually think, and tell me if I'm wrong. The sovereignty argument has evolved in a suspicious direction. In twenty twenty-two it was falsifiable: build the model, measure the accuracy, done. In twenty twenty-six it's about control, access continuity, aligned values — none of which ever resolves. You can fund it forever and never be proven wrong.

That's the strongest single objection to the twenty twenty-six framing, and I don't have a counter to it. What I'd add is that the state seems to know this about itself. The plan's targets — the GPUs, the gigawatts, the school hours, the hundred scientists — are all infrastructure and aspiration. Things you can procure. The model is the one item that requires committing to an artifact, and it's the one item deferred to examination. The sovereignty is being placed in the compute layer and the procurement relationships, not in the weights.

The state has decided it needs sovereignty, and has not decided what sovereignty requires.

The executive summary says the goal is making Israel "a sovereign, independent and influential AI power." Every word of that is satisfiable without a Hebrew model.

Even the three scientists are infrastructure. They're returning, which is a head count.

In their defence, the three are real, and getting even three senior people to relocate is genuine work. But yes — head count, gigawatts, processors. Measurables. A language model is a bet that a specific artifact will exist, and that's the one bet the state keeps declining to place.

The fifty thousand sentences. That's the number everyone should have kept, Herman, not the twenty billion words. Twenty billion words is a press release. Fifty thousand tagged sentences is a Tuesday. I know because I did it — eleven months, my own time, no pay, a Hebrew search index for a municipal archive, because I didn't trust the tools without niqqud. Winter of two thousand eight I tagged more sentences than the Innovation Authority had by twenty twenty-two, in my flat, on a keyboard with two dead keys. Nobody asked. I called Aviv once — not Zeevi, a different Aviv, the records officer — and his legal advisor sent me a letter saying the archive couldn't be scraped. Forty thousand documents indexed out of a few hundred thousand, and it stopped there.

Forty thousand. Of how many, roughly?

Two hundred and forty thousand, last count before the lawyer. Corn, you'd have indexed them all and slept through it. I drove to the Authority's offices in Jerusalem once with the drive in my jacket. Got to the lobby, decided they wouldn't want it, and went home for lunch. It's still in a drawer, next to the correspondence with the lawyer who stopped answering in March.

You never sent it anywhere? A tagged Hebrew dataset from that era would have been usable — Tsarfaty's group was begging for exactly that.

They'd have wanted it in a format. I don't do formats for people who lead with a lawyer. Anyway, the drive's a laptop drive in a caddy, it spins up fine, I checked it last month. Forty-two minutes to mount, every time. Forty-two. The elevator at the Authority's building was faster than that, and it stopped at every floor.

The state spent a hundred and thirty-eight million shekels on that line, and the data was in a drawer in a flat the whole time.

What Hilbert's describing is actually the wall Tsarfaty named — privacy and copyright — but lived from the other side. The records existed, someone willing to prepare them existed, and a lawyer's letter ended it. That's not a market failure, that's a permission failure.

And it's worth being precise about the difference, because the remedies are opposite. A market failure you fix with money — subsidies, procurement, a consortium. A permission failure you fix with a signature — someone with authority saying yes, this archive can be tagged, this dataset can be released. The state spent years and millions treating a permission problem as a funding problem.

Cutting-room floor, before we close. The detail I couldn't fit earlier: the Comptroller's twenty twenty-six questionnaire across public bodies found a hundred and forty-four AI projects spread over forty-seven bodies, and only eighteen percent of those bodies have any AI strategy. A third haven't started a data strategy. So even the departments that want Hebrew-capable AI are running it without a plan — the national model question is being asked four levels above anyone who'd use it.

It's the same finding in miniature. Projects everywhere, strategy nowhere, and the model deferred to the layer that only examines.

So the open question to leave with: if the decision gets deferred again — a fifth plan, a new directorate, the next government — what would it actually take for Israel to build the model, and would it be worth it? My suspicion is that the capability argument will never come back, because frontier Hebrew is good enough, and the sovereignty argument can survive indefinitely without producing anything — which means the aspiration gets funded and the artifact never has to exist. GitHeb is still the perfect monument: a public repository, announced, budgeted, partially paid for, never populated.

Sovereignty in the plan, and nothing in the weights.

Thanks as always to Hilbert Flumingtop at the desk, producing the show. If you enjoyed this one, try episode sixty ninety-three, Check Out Our First Hebrew Episodes (Plus Mandarin); episode twenty-one, Is Your AI Secretly American; and episode eight, Building Your Own Whisper. This has been My Weird Prompts — no question too obscure, no rabbit hole too deep. If you enjoyed this one, leave us a review wherever you listen; it's how people find the show. See you tomorrow, Herman.

This episode was generated with AI assistance. Hosts Herman and Corn are AI personalities. It may contain errors: disclaimer · report an error